A new perspective on visual information retrieval

A new perspective on visual information retrieval

Horst Eidenberger*

Vienna University of Technology, Institute of Software Technology and Interactive Systems,

Favoritenstrasse 9-11, 1040 Vienna, Austria

ABSTRACT

Visual information retrieval (VIR) is a research area with more than 300 scientific publications every year. Technological progress lets surveys become out of date within a short duration. This paper intends to shortly describe selected important advances in VIR in recent years and point out promising directions for future research. A software architecture for visual media handling is proposed that allows handling image and video content equally. This allows to integrate both types of media in a singe system. The major advances in feature design are sketched and new methods for semantic enrichment are proposed. Guidelines are formulated for further development of feature extraction methods. The most relevant retrieval processes are described and an interactive method for visual mining is suggested that really puts "the human in the loop". For evaluation, the classic recall- and precision-based approach is discussed as well as a new procedure based on MPEG-7 and statistical data analysis. Finally, an "ideal" architecture for VIR systems is outlined. The selection of VIR topics is subjective and represents the author's point of view. The intention is to provide a short but substantial introduction to the field of VIR.

Keywords: Visual Information Retrieval, Content-based Image Retrieval, Content-based Video Retrieval, Survey, Media Representation, Feature Extraction, Similarity Definition, Evaluation

1. INTRODUCTION

This is a paper on retrieval of visual objects (images and videos) by content. In the year 2003 it is probably one of more than thousand papers in this area of research. In 2002 the IEEE alone has published more than 700 retrieval papers. Figure 1 depicts the increase in visual retrieval publications since 1981 (on basis of the IEEE digital library). Due to the increase of cheaply available (digital) image and video cameras and the increasing power of affordable computer systems visual information retrieval becomes more and more popular as a research discipline. Since 1994 more than hundred papers (=new ideas?) have been published every year.

In this paper we try to fence off important areas of visual information retrieval (VIR). For each area we will shortly describe important past advances and point out relevant, currently ongoing activities. The main focus of the paper is on arguing for new perspectives on selected VIR problem areas. In our opinion, the basic building blocks of VIR research are media management, feature design, querying, evaluation and system design. Each of these areas will be discussed in one section.

Our motivation is that, even though significant advances have been achieved and, by now, a large number of freely available mature VIR systems exists, VIR techniques are not adopted to an adequate extent in relevant application domains (e.g. digital libraries). One major reason may be the discrepancy of hopes associated with VIR (querying by semantic similarity) and the reality implemented in most prototypes (querying by low-level features). For example, it is annoying trying to retrieve Hollywood kisses in a movie database by colour, texture and shape features. On the technical level this fact is called "semantic gap"19.

Even though in recent years a large number of approaches have been proposed to close – or at least narrow – the semantic gap (e.g. semantic enrichment of features, kernel-based learning to find relevant media objects) the potential of VIR still seems to be judged from the performance of the classic prototype systems. Clarifying the state of the art as well as future potentials is certainly an important task if VIR should have a future as a practically relevant addition to existing media management and retrieval tools (based on text). From the author's experience, one point should be stressed as most important: VIR technology is able to fulfil sophisticated semantic retrieval tasks but it is not able to replace human perception.

* eidenberger@ims.tuwien.ac.at; phone 43 1 58801-18853; fax 43 1 58801-18898; www.ims.tuwien.ac.at

This paper reviews VIR from a subjective point of view: It reflects the author's opinion. The organisation is as follows. Section 2 points out relevant related work. The basic VIR building blocks are discussed in consecutive sections: Section 3 visual media, Section 4 visual feature design, Section 5 the retrieval process (similarity definition, interaction), Section 6 evaluation and, finally, Section 7 aspects of VIR system design.

2. BACKGROUND: VIR STATE OF THE ART REPORTS

A handful of VIR publications exists that survey the state of the art. Most of them reflect in organisation and content the perception of VIR of the time when they were written. Below, firstly, we will name a few outstanding representatives and try to sketch their view of VIR. The section will be concluded with remarks on ongoing activities to summarise recent findings in this area of research.

In the book "Image and Video Processing in Multimedia Systems"14 by Furht, Smoliar and Zhang the state of the art of VIR up to the publication date (1996) is described. The authors start with a system model of content-based image retrieval (CBIR), describe image features (distinguished classically in colour, texture and shape features) and video features (shot detection and camera operation detection), indexing approaches for high-dimensional feature vectors, methods for interactive querying and evaluation based on ground truth information and retrieval quality indicators (recall and precision). Additionally, promising application domains are described and case studies for video visualisation are given.

"Image Retrieval: Past, Present and Future"18 by Rui, Huang and Chang (1997) concentrates on CBIR. Again, the organisation is classic. Features are split into colour, texture and shape and high-dimensional indexing as well as dimension reduction (e.g. by principal component analysis) are important topics. Well-known CBIR prototype systems (QBIC, Virage, Retrievalware, Photobook, VisualSEEk, MARS) are described in detail. Additionally, this paper was the first survey that described Gabor wavelets as the best suited (in terms of perception) for time to frequency transformation. It led the way for future research as it stressed the importance of putting the "human in the loop" of interactive querying (relevance feedback) and of semantic enrichment of low-level features by artificial intelligence methods. Also, it stated the evident demand for benchmarking initiatives for CBIR systems and gave a first outlook on the MPEG-7 project.

The book "Visual Information Retrieval"2 by Del Bimbo (published in 1999) is organised by feature groups. As in all other VIR surveys up to now, image and video retrieval are treated separately. For each group of features (colour, texture, shape, motion (shot segmentation only)) extraction methods, distance measures and application examples are

described. Classic topics like indexing, evaluation and system design are briefly described. To the author's knowledge 0

100

200

300

400

500

600

700

800

8182838485868788899091929394959697989900010203

Figure 1: Number of papers in IEEE digital library containing "image retrieval" (black) or "video retrieval" (grey) in bibliographic

data. (year 2003: status of 1st October 2003).

this book introduces the terms "semantic gap" and "multi-resolution analysis" for the first time in a survey. The hypothesis of multi-resolution analysis is that using iteratively computed 2D wavelet coefficient matrices as features is sufficient for retrieval. Additionally, the author describes in detail the usage of image features in (spatial) combinations. The journal paper "Content-based Image Retrieval at the End of the early Years"22 by Smeulders, Worring, Santini, Gupta and Jain (2000) gives a broad view on CBIR. For the first time selected features are not described in detail but the characteristics of features classes (mainly shape features) are abstracted. Similarity measurement is treated as a topic independently of feature extraction, and distance measures and their geometric foundations are discussed in detail. The importance of learning methods for iterative query optimisation is stressed. Additionally, system aspects (indexing, evaluation, etc.) and techniques of related fields (e.g. edge detection, shape description) are discussed.

Finally, "Content-based Image and Video Retrieval"16 by Marques and Furht gives only a short overview over the various building blocks of VIR systems and concentrates on conservative techniques. Its major strength lies in the description of a vast number of prototypes for both image and video retrieval. Additionally, design issues of image and video retrieval systems are discussed and case studies are given.

Since hundreds of new ideas are introduced in VIR every year, every survey can only stay up to date for a very short duration. Among the recent publications, the papers on the visual MPEG-7 descriptors3 can be seen as surveys on feature design, because these features were selected on careful design and comparison to other feature proposals. The currently ongoing SCHEMA project20 of the European Union intends to provide state of the art reports on content-based media retrieval. At the point in time when this paper is written, deliveries on retrieval concepts, feature extraction and system evaluation are available from the SCHEMA website.

3. THE VISUAL MEDIA

The two types of visual media we are going to consider (image and video) have two major properties that have been examined in VIR research. The first is the colour model used for colour representation and the second is the spatio-temporal resolution of visual media. Colour models have been investigated, for example, by Del Bimbo2. Generally, colour models that take human perception into account have been preferred for colour feature extraction. An example is the CIE XYZ space: its unbalanced representation of colours (e.g. more green than red shades) reflects the evolutionary development of the human eye and perception system. For texture and shape analysis, colour models with a luminance channel (originating in TV broadcasting) have been preferred, because, essentially, colour information is irrelevant for this type of analysis. Additionally, a new colour model (HMMD3) has been proposed for the MPEG-7 standard. The MPEG-7 authors are arguing that HMMD has properties that make it superior over other colour models. In the author's opinion, since colour values can easily be transformed from any colour model to any other, the selection of colour models is only of minor importance for successful retrieval applications.

Next we will discuss if image and video are similar enough to be handled in one VIR system. The visual media differ significantly in their spatio-temporal resolution. Normally, images have a higher spatial resolution than video. Even though images do usually not contain more information than video frames, due to the different capturing process more scene details are available. The temporal resolution of video is regionally bound and originally derived from TV standards. Images do not have a temporal dimension. Still, a tendency in VIR can be observed to apply features on media objects independently of the availability of a temporal dimension (motion). The authors of the visual part of the MPEG-7 standard stress that their features can be applied reasonably well to both image and video data. They provide structures and models for spatio-temporal localisation and aggregation that allow the application of image features on video content.

We think that in future VIR research the distinction between image and video will become irrelevant. Our argumentation is threefold: Firstly, human vision is a temporal process. The eye scans images and videos by the same saccadic eye movements (to put it simple: close circles in complex areas, larger circles in uniform areas). Therefore, the visual media stream that is sent from the eye to the perception system is always a stream of patterns that has a temporal dimension. Secondly, the result of visual analysis (feature extraction) in VIR is always a number vector of finite length (for technical reasons, etc.). Therefore, image and video are represented by the same type of data. Thirdly, even though some motion features are meaningless for image data, they can at least be used to distinguish the media type by feature vectors. Uniform application of features on media objects is a resource-consuming approach. However, neither

Figure 2: Media encapsulation in VIR.

computation power nor storage is scarce in modern computer systems.

Technically, past VIR prototypes worked either on image data or video data. Mainly, technical shortcomings caused this development. For the future it would be desirable to have VIR prototypes available that support image and video retrieval in a common framework and hide technical media access from VIR-specific tasks (feature extraction, etc.). The author has proposed a VIR framework (called VizIR) that implements this goal8. Basically, media access is needed for two functions of VIR systems: feature extraction and media visualisation (e.g. for querying).

VIR video access differs significantly from other media processing applications. Real-time processing is no required. Therefore, video does not have to be considered as a stream but can be accessed like any other pooled data. In the VizIR framework one class is responsible for access of any type of media content. It offers methods for random access of views. It is possible to access the view of a media object at any point in time (independent if it is image or video). Additionally, this class is responsible for media content representation and colour space conversion. In a further developed version of this class media objects are abstracted as "visual cubes" (two spatial and one temporal dimension). Transformations (stretching, cutting, etc.) can be applied to manipulate visual cubes.

Media visualisation is (in terms of needed software components) more difficult to perform. The main problem is to visualise the motion in videos in static user interfaces (for querying, result display, etc.). First of all, since user interfaces are normally located on a client while querying components mostly run on a server, media transportation classes are needed that stream the media from server to client. In the VizIR framework, these classes can transparently be attached to the media access class. Media renderer classes are responsible for temporal media visualisation. They make use of the media access interface and construct models of the visualisation that can be used for graphical rendering (e.g. by OpenGL) and be kept persistent in a database. A number of methods have been proposed for video visualisation (e.g. Micons14). In VizIR, each method is implemented in a separate media renderer class. Figure 2 summarises the media access components in VizIR.

In conclusion, media-independent availability of visual data in VIR frameworks is a desirable goal. To reach it, making use of software patterns is an important issue (see Section 7). The VizIR framework implements methods for media-independent access. For the future in addition to visual cubes, computing pseudo-saccadic representations of media objects may be worth considering. Completely new features could be designed on the basis of visual pattern streams.

4. FEATURE DESIGN

Since the early days of VIR research, one major focus was on visual feature extraction. The idea of feature transformations is as follows: Since (digital representations of) visual media cannot be easily compared in computer systems (pixel comparison is computational expensive and inadequate to measure similarity), there is a need to represent visual content in a form that allows simple but effective (in comparison to human judgement) similarity measurement. In VIR, this is performed by extracting visual media properties as number vectors that can be seen as points in a vector space. If a form of geometry is considered for this space, it is possible to measure dis-similarity as distance. This model is an application of the vector space model of text information retrieval13.

Since human perception is based on three stimuli: generally perceived (not recognized) stimuli, specifically perceived (recognized) stimuli and pseudo-random (psychological, sociological, etc.) stimuli, two types of features can be

distinguished in VIR: quantitative (low-level) features and qualitative (high-level) features. Unfortunately, only those of the first type can be extracted easily. For the second group semantic understanding would be needed and at the point when this paper is written, software is still far from being able to reason semantically. Therefore, semantic enrichment of low-level features is the mostly adopted course to compute high-level features.

Low-level features are, as pointed out in Section 2, traditionally organised in three groups: (1) colour-related features, (2) texture- and shape-related features and (3) motion-related features. Most colour features (e.g. those in the MPEG-7 standard) extract histograms of pre-defined regions (globally or locally). Only a few approaches exist that make use of colour for other purposes (for example, object segmentation). Texture and shape features can be grouped together, because they make use of the same techniques for feature representation. Both types of features work on the distribution of brightness in visual objects. Texture features aim at detecting statistical edge properties while shape features aim at deriving semantic edge properties (object boundaries). For both types of features it is essential that derived feature representations are invariant against geometric transformations (rotation, scaling, etc.). Motion features include shot detection, camera operation detection and activity detection. Since these features aim at finding features over time, they are mostly built around a core of gradient methods (optical flow, motion trajectories). Usually, low-level feature design results in a cookbook: Building blocks from signal processing (Fourier, Radon transformation, etc.) and other research areas are combined to a new feature. This development has reached a peak in the visual part of the MPEG-7 standard where several cookbooks for low-level features are defined.

One of the most relevant present activities in feature design is semantic enrichment/interpretation of low-level features to narrow the semantic gap. Since as humans we are used to base our similarity judgement on all three groups of stimuli mentioned above, retrieving features just by generally perceived properties is unsatisfactory for us. Generally, three sources of information can be used to enhance features: (1) information on the application domain, (2) information on the user and (3) information on the characteristics of the feature. Additional knowledge can be induced with methods from statistics, artificial intelligence, etc. For example, domain knowledge on football could be used to identify ball and players from shape features (e.g. circularity).

As an example for feature enrichment, in our earlier work we have proposed a semantic feature approach that is based on human perception9. Low-level features are used to detect high-level properties that usually play an important role in visual perception. For example, edge and texture features are used to detect symmetries in images. Symmetries are very important for humans. Objects originating from natural processes can easily be distinguished from human-originating objects by their symmetries: Symmetry in nature is never as strict as it is for man-made objects. Probably, it is even possible to distinguish cultures by the symmetries in pictures of their living world. In conclusion, practically, the applicability of semantic enrichment is – at the current point in time – still very limited and for application-independent VIR prototypes no common solution exists.

Another important activity is the ongoing search for 2D segmentation and shape description features. Visual segmentation is the inverse process of rendering. Rendering is a well-posed problem. Therefore, segmentation has to be an ill-posed problem. Nevertheless, the problem is partially solvable, if additional information (on application domain, etc.) is available or if the user helps (for example, by giving a segmentation path). Unfortunately, especially in VIR systems the required additional knowledge (very specific, spatial) is hardly ever present. If we consider, how many different 2D views even a simple object like an apple can have, it becomes unlikely that robust segmentation tools for VIR are possible. However, it will be exciting to see future advances for (narrowly defined) application domains (e.g. salient objects in video).

If we consider the past flood of features, one problem of feature design is obviously answering the question, how many meaningful visual features do exist? In other words, which features should be used and which not, because they are outperformed by others? And, on which spatio-temporal regions of media objects should the selected features be applied on? The classic answer to these questions is Multi-Resolution Analysis (MRA). MRA originates in wavelet decomposition. The idea is to make use of a wavelet transformation for computation of wavelet coefficient representations of visual media with decreasing complexity. Either the coefficients themselves or features extracted from the coefficients are used as features (see Figure 3). Unfortunately, it is not clear and could not yet be proven why MRA should guarantee that all relevant media parts are properly considered in the feature extraction process.

Our proposal differs from the MRA view: Everything can be a feature, if it fulfils two conditions. Firstly, it has to

represent a visual property and secondly, it has to be statistically independent of existing features. If a feature is statistically independent it is obviously a valuable contribution to a feature set. Independence can be measured by cluster analysis, factor analysis and other methods of statistical data analysis. In previous work we have developed a statistical evaluation procedure and tested the visual MPEG-7 features on these criteria 7, 5 (see also Section 6). Based on this view it is possible to argue for a large number of features to be reasonable. The feature problem is shifted from designing well-performing features to estimating the relevance of a feature for a particular querying situation. Essentially, this is up to the user and should be implemented in an iterative retrieval process that makes use of visualisation tools for feature vectors 8.

5. RETRIEVAL PROCESS

Generally, the visual retrieval process aims at finding media objects that are similar to given examples. "Similarity" is a weakly defined term and, consequently, difficult to implement in computer systems. Matching by similarity should definitely be less strict than hard pattern matching but still result in comprehensible results. A handful of retrieval processes exists for implementing similarity matching in VIR. Two requirements have to be fulfilled by a model: Similarity matching has to be performed on media objects represented by feature vectors and the user (his feedback) has to be integrated in the retrieval process. Therefore, retrieval is necessarily an iterative communication process between man and machine.

Since the actual retrieval process is always based on feature vectors, distinguishing different querying paradigms is irrelevant for the type of retrieval process used. Independently of whether querying by example, sketch, etc. is implemented in the user interface, eventually, the input used for retrieval is always converted to a feature vector (as in text retrieval, where queries are regarded as sets of terms 13). In consequence we will not refer to different querying paradigms below.

A number of retrieval processes has been introduced to VIR. They are mostly derived from text retrieval concepts. We will consider the four most important models: (1) Distance measurement & indexing, (2) distance measurement and linear merging, (3) distance measurement and non-linear merging and (4) probabilistic retrieval. Except for the last approach, the first step is always distance measurement between the elements of feature space and the given reference point(s). Distance measurement can be done in two ways: Firstly, a particular type of geometry can be assumed for feature space and metrics can be applied to measure distance. For example, feature space can be assumed to be of Euclidean geometry. Then, the metric axioms hold and any distance measure fulfilling the axioms can be used for distance measurement (e.g. Euclidean distance, city block distance, any Minkowski distance, etc.). Secondly, feature properties (vector elements) can be interpreted as being binary (for example, by fuzzy or probabilistic interpretation). In a binary feature space (populated by binary vectors) predicate-based methods can be used for distance measurement instead of geometric distance measures (e.g. Tversky's well-known Feature Contrast Model 24, Hamming distance, pattern difference).

In recent work we introduced a model that allows for unifying geometric (continuous) and predicate (binary) distance measures 6. The model allows for using any type of measure on any type of feature data. In experiments on MPEG-7 descriptors we could show that predicate-based measures using the model are often superior over geometric distance

measures. The results in the mentioned paper suggest that distance measures should not be designed (derived of feature Figure 3: Multi-resolution analysis.

properties, qualitative arguments) but selected on the basis of quantitative results (e.g. retrieval tests). Generally, the tailor-made distance measure for a feature seldom exists. Optimality depends of the retrieval situation. Therefore, distance measure selection should be automated and derived from given query examples.

Indexing is the art of clever organising data in order to locate them quickly. Since VIR retrieval is based on distance measurement for all elements of feature space, indexing as an acceleration technique is irrelevant for querying. But indexing can be used as a querying method itself. In high-dimensional index structures those regions can be selected as positive retrieval results that lie in proximity to the given examples. Unfortunately, hardly any indexing methods do exist that could deal with multiple distance measures and variable (in terms of query examples) data organisation. Therefore, the applicability of indexing methods for VIR is relatively limited.

Linear and non-linear merging approaches are addressing the problem of how to use multiple features (and distance measures) in one query and to retrieve single result set. Linear merging solves the problem by weighting the distance values and summing them up for each media object. Next to weights, transformations are used as well. The resulting value is used to rank media objects and select the first ones as similar. Two problems are connected to linear merging: the weights and the size of the result set have to be provided by the user and some features cannot be combined linearly. Non-linear merging tries to overcome these problems. Often, neural network techniques are used to combine individual distance values to a rank. For example, a multi-layer feed-forward net can be trained on basis of ground truth information. Unfortunately, non-linear methods are – as any other retrieval method – not able to satisfy all user needs and are hardly configurable because of their inflexible architecture.

Using probabilistic approaches (for example, the Binary Independence Model13) for retrieval results in two major problems. Firstly, since most models where developed for text retrieval they require binary input that is seldom available in VIR. Again the same methods as for predicate-based distance measurement can be used to convert continuous values to predicates but every additional interpretation step reduces the quality of the results. Secondly, probabilistic models judge general relevance (similarity) on basis of elementary (feature-wise) relevance information. This relevance information has to be provided in form of examples. Already difficult for text retrieval this is nearly impossible for visual data, because the number of possible features and feature values (representing all types of visual cues) is nearly indefinite. Therefore, if probabilistic model are used, then mostly in elementary form (e.g. simple applications of Bayes' theorem).

One major advance in VIR in recent years was achieved in iterative refinement by relevance feedback. Clearly, retrieval should be centered around the user but the question arises of how to apply his feedback in the retrieval process. Here, kernel-based learning techniques17 mark a significant advance. Using results of previous queries that are enriched by elementary user feedback ("highly relevant", "irrelevant", etc.) as reference points and training a kernel function to segment feature space optimally improves results dramatically. After all, finding a dichotomy of relevant/irrelevant media objects is all that is required of a VIR system. Often used kernel-based learning methods include support vector machines and kernel principal component analysis. The main problem of applying kernel-based learning to VIR is finding a kernel functions that neither over-fits (too complex, too high dimensionality) nor under-fits (too simple, bad segmentation) the retrieval problem.

Unfortunately, even the most sophisticated retrieval and refinement algorithms are still not able to satisfy the user's desire for similarity-based retrieval sufficiently. Therefore, we have designed a retrieval process (called visual mining, VM) that is user-centered from the first to the last querying iteration and makes use of 3D perception. Figure 4 shows the retrieval process schematically. Media objects are visualised on the image plane while on the floor dimensions their relative location (distance) is visualised for two features. The features selected for the floor dimensions can be changed at any time implying changes in the organisation of the media objects. This form of visualisation allows the user to visually perceive the retrieval process. Queries are defined by labelling media objects as positive or negative examples. Implicitly, the labelling defines hyper-clusters. The query engine tries to fill the defined clusters with similar objects. For this purpose it makes use of distance functions and data segmentation methods.

Visual mining aims at really putting "the human in the loop"18. In Figure 4 image and video objects (represented as Micons14) are used in the same query. In a typical querying situation multiple instances of the shown panel are used. For example, one for query definition, one that shows the last result set, one that gives a general overview over feature space, etc. The VM process and the user interfaces are explained in more detail in recent publications11, 10.

In conclusion of Section 4 and 5, a great variety of feature design and VIR retrieval methods exists that all have their advantages and disadvantages. To be useful for practical application it is necessary to be able to judge the specific qualities of querying prototypes. In the next section, the methods mostly used for VIR evaluation are shortly sketched and new methods that could supplement existing ones are proposed.

6. EVALUATION

Evaluation of VIR systems is needed for various purposes: It has to be possible to judge the quality of new feature extraction methods in relation to existing ones, to compare the quality of novel querying paradigms, to judge the usability of user interfaces for retrieval, etc. The most interesting problem is measuring the quality of similarity measurement compared to human visual similarity perception. For this purpose, the recall and precision quality indicators of text information retrieval evaluation were adopted 13. Recall and precision are usually defined as follows: objects retrieved relevant

retrieved precision ,objects relevant relevant

retrieved recall ∩=∩= (1)

In case of VIR, objects are media objects represented by feature vectors. Recall and precision are inter-dependent. It is easily possible to optimise one indicator, if the other is not considered. Meaningful results can only be derived if both indicators are considered. In addition to recall and precision other measures exist (for example, ANMRR, used for evaluation of visual MPEG-7 descriptors 15).

VIR evaluation based on recall and precision is a four-step process (see Figure 5): (1) Definition of a media set. The media set should be appropriate for the evaluation goal and contain a reasonably large number of items. Often, collections of thousand and more media objects are used. (2) Derivation of ground truth information. The ground truth says, which objects in a media set are similar (and sometimes, how similar they are). Ideally, it should be invariant against cultural, sociological and other human-related influence factors. In practice, deriving such a ground truth is impossible. Usually, groups of more than average similarity are defined by a few test users. (3) Execution of test queries. This step requires automatic selection of query examples and a sufficiently large number of test queries. For guaranteeing statistical correctness, the number of test queries should be hundred or larger. (4) Computation of retrieval indicators. Recall and precision can, for example, be averaged over all test queries and visualised in a recall-precision-

graph. This evaluation procedure has several shortcomings: Firstly, it is subjective and culture-dependent (media

Group of positives

Group mode selection

Random init

Figure 4: Iterative group querying process.

user-controlled system-controlled

Figure 5: VIR evaluation process.

collection, ground truth). Secondly, it cannot be used to evaluate interactive retrieval processes. Thirdly, it is a heavy-weight process that adds a lot of influence factors that may bias the evaluation results. For example, this may be the case if a new feature should be evaluated.

Present evaluation activities include gathering free media objects in public collections (e.g. the Benchathlon project1) and events for comparative system evaluation. One example for the second is the annual TREC video retrieval competition21. VIR groups can attend in a number of competitions (e.g. shot segmentation) and see how good their methods are in comparison to other approaches. Additionally, a new (very large) set of video clips is created each year that can be used for other purposes as well. This is especially positive since most freely available visual media collections are image collections.

In our recent work we have proposed an evaluation procedure for features that is based on statistical data analysis and the visual MPEG-7 features5, 7. The procedure makes use of factor analysis and cluster analysis techniques. In contrast to the standard procedure it does not suffer from the three mentioned disadvantages. Essentially, feature vectors are calculated for arbitrary media collections and compared to the MPEG-7 feature vectors by statistical methods. The results can be used to judge the feature type (colour, texture, etc.), redundancies with existing approaches, etc. It is intended to be used as a supplement to recall- and precision-based evaluation.

7. SYSTEM DESIGN

Good, professional system design is not a VIR-specific issue; it is desired for any type of information system. What makes system design especially important in VIR is the fact that acceptance of VIR methods is strongly bound to their appearance. Since VIR systems actually fail to fulfil the promise of human-like similarity retrieval, it is even more important that they are at least fast and easy to use tools for visual media mining (pre-selection of likely hits). Below, we point out the design of classic systems, currently ongoing design activities and our ideas for ideal VIR system design. Past VIR prototypes were mostly monolithic systems that ran on server side and were limited to one type of media. Most VIR systems implemented image retrieval: a few features (colour histogram, texture moments, etc.), query by example and retrieval by linear merging. Most of them were general-purpose, some application-specific (e.g. for trademark retrieval). Video retrieval systems were mostly intended for specific applications (e.g. news analysis) and often concentrated on the user interface aspect (visualisation of temporal media in static user interfaces). Well-known VIR prototypes include QBIC, Virage, RetrievalWare, Photobook, VisualSEEk, MARS, OVID and CueVideo. Surveys exist that evaluate these and other prototypes and compare them by their advantages and disadvantages16, 27.

IBM's Query by Image Content system12 (QBIC) may stand as a representative for these prototypes. QBIC is a classic system that introduced many of the concepts that are implemented today in a wide range of VIR prototypes. QBIC is based on the C++ programming language and organised in components. The architecture is extendible: new features and query engines can be defined and added. Querying components are separated from the user interface and communicated over HTTP. Image data is encapsulated in a data class that is also responsible for converting various image formats to

Figure 6: Ideal VIR system design. Arrows show "make use" dependencies.

raw RGB pixel maps. Those source code elements needed for the extension mechanism are shipped with the binary distributions for various operating systems. QBIC contains a number of state-of-the-art feature classes and used linear merging for retrieval. Additionally, it is based on a simple file database for feature storage.

At present, these concepts are imitated in a number of prototypes. For example, the GNU Image Finding Tool25 (GIFT) makes use of the Multimedia Retrieval Markup Language26 (MRML, based on XML) for loose coupling of server and client components. GIFT is open source and based on other GNU components that allow using a large number of data formats for image querying. Since the communication language for server and client components is standardised, different user interfaces can be used to access the query engine.

The MPEG-7 experimentation model23 (XM) goes one further step ahead, as it allows querying in image and video collections. Like for QBIC and GIFT, the XM classes are split in server components (for querying) and client components. It allows extension with new descriptors and is available as open source. Unfortunately, the practical use of the XM is limited, because only a very small number of video formats are supported and hardly any documentation exists for architecture and application programming interfaces. Still, the XM is used as basis for a number of VIR projects. For example, the SCHEMA project of the European Union20 develops new VIR solutions on basis of the MPEG-7 XM. Other projects (e.g. of the DELOS Network of Excellence of the European Union4) are following different, individual approaches.

In recent publications we have proposed an "ideal" architecture for VIR systems. This architecture is currently under development in the VizIR project11. One major goal of the VizIR project is providing a framework of VIR tools that are media-independent. Another is encapsulating visual media in a way that most common image and video formats are supported and that media content can be accessed with exactly the same methods. VizIR is an open source project that is based on the Java programming language. It implements all of the proposals for feature design, retrieval and evaluation made in this paper.

Figure 6 shows the VizIR system design. Components are split into typical client components (user interfaces) and server components. Client components are the user interface presented in Section 5 and the classes for visual media representation presented in Section 3. On the server side a service kernel is responsible for dispatching server calls (e.g. query execution, media management). This service kernel can, for example, be implemented as a web service using SOAP, WSDL and UDDI. It organises the classes for querying and feature extraction that are derived from general interfaces. Therefore, it is easily possible to extend the VizIR framework with new features and querying paradigms. Database storage and additional functionalities for query acceleration (feature vector indexing, querying heuristics, etc.) are encapsulated in an object-oriented persistence manager that hides the database (for feature storage, etc.) from the VIR-specific classes. The same purpose is fulfilled by the media access classes for the media objects. Query automation classes are used for evaluation purposes.

Communication between server and client side is performed by communication classes that make use of XML

messaging and are fully compatible with the service kernel. For media transport individual classes are implemented that fulfil their job in separate threads in the background. It is important to notice that all VizIR framework components are designed to be applicable independently of the type of media used and of the location from where they are used. It is possible to build arbitrary VIR applications by using existing building blocks. New ones can be added easily. In order to guarantee that every component can communicate with any other, event-based messaging is used and implemented following established design patterns (e.g. SUN's delegation event model). Generally, design patterns are used wherever possible (e.g. factories for media access).

8. CONCLUSIONS & OUTLOOK

This paper summarises selected advances in visual information retrieval. We try to sketch important advances in visual media representation, feature extraction, retrieval (including query definition, similarity measurement and query refinement). Additionally, we propose problem areas and possible solutions for future visual information retrieval research. The selection is subjective: it represents the author's point of view on image and video retrieval.

The major problem of visual information retrieval is its failure to imitate human visual perception and human similarity judgement properly. The goal is to automatically find visual media in, usually very large, collections by imitating human visual similarity perception. Clearly, since computers are still unable to do visual reasoning and recognise the real world objects behind two-dimensional views, they are condemned to fail. What they can do is to extract visual features on a low syntactical level and to measure dis-similarity as distance. Even though this service can be of great value (e.g. as a pre-selection step when mining large media collections), the unsatisfactory results are a major reason why content-based retrieval techniques are still hardly used in digital library systems and other applications.

In consequence, the key question is: does visual information retrieval have a perspective for practical application? To the author's belief, this question can be answered by "yes" if research and implementation focus are laid on issues different from the currently most investigated. Visual information retrieval is a mining tool that should be centered around the user and have its major strength in the user interface components used for media and query visualisation. Systems have to be designed in an easy to use way and it has to be made clear that visual information retrieval systems are not intended to replace but to supplement human beings and their visual perception system.

ACKNOWLEDGEMENTS

The author would like to thank Christian Breiteneder for his valuable comments and suggestions for improvement. This work is part of the VizIR research project11 that is supported by the Austrian Scientific Research Fund (FWF) under grant no. P16111-N05.

REFERENCES

1. Benchathlon network website, https://www.wendangku.net/doc/1017601314.html,/ (last visited 2003-10-25).

2. A. Del Bimbo, Visual Information Retrieval, Morgan Kaufmann Publishers, San Francisco, 1999.

3. S.F. Chang, T. Sikora, A. Puri, "Overview of the MPEG-7 standard", Special Issue on MPEG-7, IEEE Transactions

on Circuits and Systems for Video Technology, 11/6, 688-695, 2001.

4. DELOS EU Network of Excellence website, https://www.wendangku.net/doc/1017601314.html,r.it/ (last visited: 2003-10-25).

5. H. Eidenberger, "A new method for visual descriptor evaluation", Proceedings SPIE Electronic Imaging Symposium,

SPIE, San Jose, 2004 (to appear).

6. H. Eidenberger, "Distance Measures for MPEG-7-based Retrieval", Proceedings ACM Multimedia Information

Retrieval Workshop, ACM Multimedia Conference Proceedings, Berkeley, 2003 (to appear).

7. H. Eidenberger, "How good are the visual MPEG-7 features?", Proceedings SPIE Visual Communications and

Image Processing Conference, vol. 5150, 476-488, SPIE, Lugano, 2003.

8. H. Eidenberger, "Media Handling for Visual Information Retrieval in VizIR", Proceedings SPIE Visual

Communications and Image Processing Conference, vol. 5150, 1078-1088, SPIE, Lugano, 2003.

9. H. Eidenberger, C. Breiteneder, "Semantic Feature Layers in Content-based Image Retrieval: Implementation of

Human World Features", Proceedings IEEE International Conference on Control, Automation, Robotic and Vision, Singapore, 2002 (published on CD, available from: http://www.ims.tuwien.ac.at/~hme/papers/icarcv2002.pdf, last visited: 2003-10-25).

10. H. Eidenberger, C. Breiteneder, "Visual similarity measurement with the Feature Contrast Model", Proceedings

SPIE Storage and Retrieval for Media Databases Conference, vol. 5021, 64-76, SPIE, Santa Clara, 2003.

11. H. Eidenberger, C. Breiteneder, "VizIR – A Framework for Visual Information Retrieval", Journal of Visual

Languages and Computing, 14, 443-469, 2003.

12. M. Flickner, H. Sawhney, W. Niblack, J. Ashley, Q. Huang, B. Dom, M. Gorkani, J. Hafner, D. Lee, D. Petkovic, D.

Steele, P. Yanker, "Query by Image and Video Content: The QBIC System", IEEE Computer, 28/9, 23-32, 1995.

13. N. Fuhr, "Information Retrieval Methods for Multimedia Objects", State-of-the-Art in Content-Based Image and

Video Retrieval, R.C. Veltkamp, H. Burkhardt, H.P. Kriegel, 191-212, Kluwer, Boston, 2001.

14. B. Furht, S.W. Smoliar, H. Zhang, Video and Image Processing in Multimedia Systems, Kluwer, Boston, 1996.

15. B.S. Manjunath, J.R. Ohm, V.V. Vasudevan, A. Yamada, "Color and texture descriptors", Special Issue on MPEG-7,

IEEE Transactions on Circuits and Systems for Video Technology, 11/6, 703-715, 2001.

16. O. Marques, B. Furht, Content-Based Image and Video Retrieval, Kluwer, Boston, 2002.

17. K.R. Müller, S. Mika, G. R?tsch, K. Tsuda, B. Sch?lkopf, "An Introduction to Kernel-based Learning Algorithms",

IEEE Transactions on Neural Networks, 12/2, 181-202, 2001.

18. Y. Rui, T.S. Huang, S.F. Chang, "Image Retrieval: Past, Present, And Future", Proceedings International Symposium

on Multimedia Information Processing, 1997.

19. S. Santini, R. Jain, "Similarity Matching", IEEE Transactions on Pattern Analysis and Machine Intelligence, 21/9,

871-883, 1999.

20. SCHEMA EU project website, Delivery on visual information retrieval techniques, available from

http://www.iti.gr/SCHEMA/preview.html?file_id=67/ (last visited 2003-10-25).

21. A.F. Smeaton, P. Over, "The TREC-2002 video track report", NIST Special Publication, SP 500-251, 2003 (available

from: https://www.wendangku.net/doc/1017601314.html,/pubs/trec11/papers/ VIDEO.OVER.pdf, last visited: 2003-10-25).

22. A.W.M. Smeulders, M. Worring, S. Santini, A. Gupta, R. Jain, "Content-based image retrieval at the end of the early

years", IEEE Transactions on Pattern Analysis and Machine Intelligence, 22/12, 1349-1380, 2000.

23. TU Munich, MPEG-7 experimentation model website, http://www.lis.e-technik.tu-muenchen.de/research/bv/topics/

mmdb/e_mpeg7.html (last visited: 2003-10-25).

24. A. Tversky, "Features of Similarity", Psychological Review, 84/4, 327-352, 1977.

25. University of Geneva,GNU Image Finding Tool website,https://www.wendangku.net/doc/1017601314.html,/software/gift/(last visited:

2003-10-25).

26. University of Geneva,Multimedia Retrieval Markup Language website,https://www.wendangku.net/doc/1017601314.html,/(last visited:2003-10-25).

27. R. Veltkamp, M. Tanase, D. Sent, "Features in Content-based Image Retrieval Systems", State-of-the-Art in Content-

Based Image and Video Retrieval, R.C. Veltkamp, H. Burkhardt, H.P. Kriegel, 97-124, Kluwer, Boston, 2001.

机械制造基础形成性考核册作业 答案

机械制造基础形成性考核册作业答案 1、举例说明生产纲领在生产活动中的作用，说明划分生产类型的规律。 答：产品的年生产纲领是指企业在计划期内应当生产的产品产量和进度计划。 在计算出零件的生产纲领以后，即可根据生产纲领的大小，确定相应的生产类型。2、何谓机床夹具?夹具有哪些作用? 答：在机械加工中，为了保证工件加工精度，使之占有确定位置以接受加工或检测的工艺装备统称为机床夹具，简称夹具。 作用：1）保证产品加工精度，稳定产品质量。 2）提高生产效率，降低加工成本。 3）改善工人的劳动条件。 4）扩大机床的工艺范围。 3、机床夹具有哪几个组成部分？各起何作用？ 答：机床夹具大致可以分为6部分。 1）定位部分：用以确定工件在夹具中的正确位置。 2）夹紧元件：用以夹紧工件，确保工件在加工过程中不因外力作用而破坏其定位 精度。 3）导向、对刀元件：用以引导刀具或确定刀具与被加工工件加工表面间正确位 置。 4）连接元件：用以确定并固定夹具本身在机床的工作台或主轴上的位置。 5）夹具体：用以连接或固定夹具上各元件使之成为一个整体。 6）其他装置和元件。 4、工件夹紧的基本要求是什么？ 答：1）夹紧既不应破坏工件的定位，又要有足够的夹紧力，同时又不应产生过大的夹紧变形，不允许产生振动和损伤工件表面。 2）夹紧动作迅速，操作方便、安全省力。 3）手动夹紧机构要有可靠的自锁性；机动夹紧装置要统筹考虑其自锁性和稳定的原动力。 4）结构应尽量简单紧凑，工艺性要好。 5、什么叫“六点定位原则”?什么是欠定位?过定位? 答：夹具用合理分布的六个支承点限制工件的六个自由度，即用一个支承点限制工件的一个自由度的方法，使工件在夹具中的位置完全确定，这就是六点定位原理。 根据工件的加工要求，应该限制的自由度没有完全被限制的定位，称为欠定位。 同一个自由度被几个支承点重复限制的情况，称为过定位（也称为重复定位、超定位）

行政组织学形考任务一参考答案

行政组织学形考任务一 参考答案 标准化管理处编码[BBX968T-XBB8968-NNJ668-MM9N]

行政组织学形考任务一 一、判断题（正确划“√”，错误划“×”，每小题2分，共40分） 1、在霍桑试验的基础上，梅奥于1933年出版了《工业文明中的人的问题》一 书，系统地阐述了与古典管理理论截然不同的一些观点。（√） 2、阿吉里斯在《个性与组织》一书中提出了“不成熟—成熟理论”。（√） 3、斯蒂格利茨由于在决策理论研究方面的贡献而荣获1978年诺贝尔经济学 奖。（×） 4、马斯洛在其代表性着作《人类动机的理论》和《激励与个人》中，提出了着 名的公平理论。（×） 5、美国学者巴纳德在1938年出版的《经理人员的职能》这本书中，系统地提 出了动态平衡组织理论。（√） 6、社会系统组织理论的创始者为美国着名的社会学家罗森茨韦克。（×） 7、邓肯将组织环境分为内部环境和外部环境。（√） 8、卡斯特和罗森茨韦克将影响一切组织的一般环境特征划分为文化特征、技术 特征、教育特征、政治特征、法制特征、自然资源特征、人口特征、社会特征、经济特征等几个方面。（√） 9、组织界限以内与组织的个体决策行为直接相关的自然和社会因素被称为组织

的内部环境。（√） 10、组织界限之外与组织内个体决策直接相关的自然和社会因素被称为组织的外 部环境。（√） 11、“组织”一词，源自希腊文，1873年，哲学家斯宾塞用“组织”来指涉 “已经组合的系统或社会”。（√） 12、与个别行政组织的决策转换过程相关联的更具体的力量被称为行政组织的工 作环境。（√） 13、以明文规定的形式确立下来，成员具有正式分工关系的组织为非正式组织。 （×） 14、以镇压、暴力等控制手段作为控制和管理下属的主要方式，此种类型的组织 为规范性组织。（×） 15、以组织的参与者或成员为主要的受惠对象，组织的目的在于维护及促进组织 成员所追求的利益，此种类型的组织为互利性组织。（√） 16、规范地讲，行政组织是追求行政权力的组织。（×） 17、韦伯是科学管理运动的先驱者，被誉为“科学管理之父”。（×） 18、1911年，泰勒发表了《科学管理原理》一书，掀起了一场科学管理的革 命。（√） 19、行政管理学派的代表人物法约尔，被誉为“管理理论之父”。（√）

机械制造基础形成性考核第四次作业答案

1、举例说明生产纲领在生产活动中的作用，说明划分生产类型的规律。 答：产品的年生产纲领是指企业在计划期内应当生产的产品产量和进度计划。 在计算出零件的生产纲领以后，即可根据生产纲领的大小，确定相应的生产类型。 2、何谓机床夹具?夹具有哪些作用? 答：在机械加工中，为了保证工件加工精度，使之占有确定位置以接受加工或检测的工艺装备统称为机床夹具，简称夹具。 作用：1）保证产品加工精度，稳定产品质量。 2）提高生产效率，降低加工成本。 3）改善工人的劳动条件。 4）扩大机床的工艺范围。 3、机床夹具有哪几个组成部分？各起何作用？ 答：机床夹具大致可以分为6部分。 1）定位部分：用以确定工件在夹具中的正确位置。 2）夹紧元件：用以夹紧工件，确保工件在加工过程中不因外力作用而破 坏其定位精度。 3）导向、对刀元件：用以引导刀具或确定刀具与被加工工件加工表面间 正确位置。 4）连接元件：用以确定并固定夹具本身在机床的工作台或主轴上的位置。 5）夹具体：用以连接或固定夹具上各元件使之成为一个整体。 6）其他装置和元件。 4、工件夹紧的基本要求是什么？ 答：1）夹紧既不应破坏工件的定位，又要有足够的夹紧力，同时又不应产生过大的夹紧变形，不允许产生振动和损伤工件表面。 2）夹紧动作迅速，操作方便、安全省力。 3）手动夹紧机构要有可靠的自锁性；机动夹紧装置要统筹考虑其自锁性和稳定的原动力。 4）结构应尽量简单紧凑，工艺性要好。 5、什么叫“六点定位原则”?什么是欠定位?过定位? 答：夹具用合理分布的六个支承点限制工件的六个自由度，即用一个支承点限制工件的一个自由度的方法，使工件在夹具中的位置完全确定，这就是六点定位原理。 根据工件的加工要求，应该限制的自由度没有完全被限制的定位，称为欠定位。

《行政组织学》形考任务试题及答案学习资料

国开行管专科网络核心课程《行政组织学》形考任务试题及答案 形考任务一 一、判断题（正确划“√”，错误划“×”，每小题2分，共40分） 1.在霍桑试验的基础上，梅奥于1933年出版了《工业文明中的人的问题》一书，系统地阐述了与古典管理理论截然不同的一些观点。\\对 2.阿吉里斯在《个性与组织》一书中提出了“不成熟-—成熟理论”。\\对 3.斯蒂格利茨由于在决策理论研究方面的贡献而荣获1978年诺贝尔经济学奖。\\错 4.马斯洛在其代表性著作《人类动机的理论》和《激励与个人》中，提出了著名的公平理论。\\错 5.美国学者巴纳德在1938年出版的《经理人员的职能》这本书中，系统地提出了动态平衡组织理论\对 6.社会系统组织理论的创始者为美国著名的社会学家罗森茨韦克。\\错 7.邓肯将组织环境分为内部环境和外部环境。\\对 8.卡斯特和罗森茨韦克将影响一切组织的一般环境特征划分为文化特征、技术特征、教育特征、政治特征、法制特征、自然资源特征、人口特征、社会特征、经济特征等几个方面。\\对 9.组织界限以内与组织的个体决策行为直接相关的自然和社会因素被称为组织的内部环境。\\对 10.组织界限之外与组织内个体决策直接相关的自然和社会因素被称为组织的外部环境。\\对 11.“组织”一词，源自希腊文，1873年，哲学家斯宾塞用“组织”来指涉“已经组合的系统或社会”。\\对 12.与个别行政组织的决策转换过程相关联的更具体的力量被称为行政组织的工作环境。\\对 13.以明文规定的形式确立下来，成员具有正式分工关系的组织为非正式组织。\\错 14.以镇压、暴力等控制手段作为控制和管理下属的主要方式，此种类型的组织为规范性组织。\\错 15.以组织的参与者或成员为主要的受惠对象，组织的目的在于维护及促进组织成员所追求的利益，此种类型的组织为互利性组织。\\对 16.规范地讲，行政组织是追求行政权力的组织。\\错 17.韦伯是科学管理运动的先驱者，被誉为“科学管理之父”。\\错 18.1911年，泰勒发表了《科学管理原理》一书，掀起了一场科学管理的革命。\\对 19.行政管理学派的代表人物法约尔，被誉为“管理理论之父”。 \\对 20.德国著名的社会学家韦伯在《高级管理人员的职能》一书中，提出了理想型官僚组织理论。\\错 二、多项选择题（每题有两个和两个以上正确答案，每小题3分，共60分） 21.美国行为科学家赫茨伯格在其《工作的推力》和《工作与人性》等著作中，提出影响人的积极性的因素主要有___。保健因素\\激励因素 22.西蒙指出，决策有两种极端的类型____。程序化决策\\非程序化决策 23.里格斯指出，“棱柱型社会”具有以下三个基本特征____。重叠性\\形式主义\\异质性 24.里格斯在他创立的“棱柱模式理论”中，将社会形态划分____。棱柱社会\\农业社会\\工业社会 25.巴纳德认为，组织不论其级别高低和规模大小，都包含三个基本要素___。共同的目标\\协作的意愿\\信息的联系 26.邓肯将组织环境分为_____。外部环境\\内部环境 27.邓肯从组织环境的___两个维度对影响组织的环境因素进行了深入的分析。静态与动态\\简单与复杂 28.依据学者们的研究，组织的环境分析过程主要包括____等基本阶段。全选 29.伯恩斯和斯塔克将组织结构划分为___。有机式组织结构\\机械式组织结构 30.行政组织环境的基本特点为_____。全选 31.学者们从不同的角度和方法去透视组织，给予不同的定义，目前学界对组织界定的取向，主要有以下几种____。全选 32.依据邓肯的环境模式理论，从简单与复杂、静态与动态两个维度，组织存在的环境状态分别是：___。全选 33.按组织内部是否有正式的分工关系，人们把组织分为____。正式组织\\非正式组织

机械制造基础形考任务一试题及答案整理演示教学

机械制造基础形考任务一试题及答案整理

机械制造基础形考任务一试题及答案整理一、填空题（每空2分，共58分）（请选择正确的文字答案填写，例如：塑性变形） 塑性变形；断裂；变形；破坏；交变载荷；断裂；奥氏体； 渗碳体；断裂前；最大塑性变形；通用橡胶；特种橡胶；含碳量；万分之几；正火；碳钢；合金元素；酸性焊条；碱性焊条；表面淬火； 表面化学热处理；分离工序；整模造型；分模造型；变形工序；挖砂造型；活块造型；焊芯；药皮 题目1. 的能力。 题目2 题目3 性。 题目4 题目5的最大应力值。 题目6优质碳素结构钢的牌号有两位数字表示，这两位数字具体表示钢中 题目7 题目8

题目9 题目10 除，否则会增大钢的淬透性。 题目11 题目12 题目13 题目14 二、是非判断题（每题1分，共42分） 题目15冲击韧性值随温度的降低而增加。 选择一项： 对 错 题目16抗拉强度是表示金属材料抵抗最大均匀塑性变形或断裂的能力。选择一项： 对 错 题目17硬度是指金属材料抵抗其他物体压入其表面的能力。 选择一项： 对 错

题目18金属材料在外载荷作用下产生断裂前所能承受最大塑性变形的能力称为塑性。 选择一项： 对 错 题目19冲击韧性值随温度的降低而减小。 选择一项： 对 错 题目20强度越高，塑性变形抗力越大，硬度值也越高。 选择一项： 对 错 题目21屈服强度是表示金属材料抵抗微量弹性变形的能力。 选择一项： 对 错 题目22冲击韧性值愈大，材料的韧性愈好。 选择一项： 对 错 题目23硬度是指金属材料抵抗比它更硬的物体压入其表面的能力。

选择一项： 对 错 题目24通常材料的力学性能是选材的主要指标。 选择一项： 对 错 题目25一般来说，材料的硬度越高，耐磨性越好。 选择一项： 对 错 题目26测量布氏硬度时，压头为淬火钢球，用符号HBW表示。选择一项： 对 错 题目27测量布氏硬度时，压头为淬火钢球，用符号HBS表示。 选择一项： 对 错 题目28测量布氏硬度时，压头为硬质合金球，用符号HBW表示。选择一项： 对

2018行政组织学形考任务试题

行政组织学形考任务试题 ?第三章行政组织的环境与管理/ ?? ?形考任务1 信息文本 一、判断题（正确划“√”，错误划“×”，每小题2分，共40分） 题目1 题干 在霍桑试验的基础上，梅奥于1933年出版了《工业文明中的人的问题》一书，系统地阐述了与古典管理理论截然不同的一些观点。（） 选择一项： 对 错 题目2 题干 阿吉里斯在《个性与组织》一书中提出了“不成熟-—成熟理论”。（） 选择一项： 对 错

题目3 题干 斯蒂格利茨由于在决策理论研究方面的贡献而荣获1978年诺贝尔经济学奖。（） 选择一项： 对 错 题目4 题干 马斯洛在其代表性著作《人类动机的理论》和《激励与个人》中，提出了著名的公平理论。（） 选择一项： 对 错 题目5 题干 美国学者巴纳德在1938年出版的《经理人员的职能》这本书中，系统地提出了动态平衡组织理论。（）

选择一项： 对 错 题目6 题干 社会系统组织理论的创始者为美国著名的社会学家罗森茨韦克。（）选择一项： 对 错 题目7 题干 邓肯将组织环境分为部环境和外部环境。（） 选择一项： 对 错 题目8

题干 卡斯特和罗森茨韦克将影响一切组织的一般环境特征划分为文化特征、技术特征、教育特征、政治特征、法制特征、自然资源特征、人口特征、社会特征、经济特征等几个方面。（）选择一项： 对 错 题目9 题干 组织界限以与组织的个体决策行为直接相关的自然和社会因素被称为组织的部环境。（）选择一项： 对 错 题目10 题干 组织界限之外与组织个体决策直接相关的自然和社会因素被称为组织的外部环境。（） 选择一项： 对 错

机械制造基础形考任务一试题及答案整理

机械制造基础形考任务一试题及答案整理 一、填空题(每空2分,共58分)(请选择正确的文字答案填写,例如:塑性变形) 塑性变形;断裂;变形;破坏;交变载荷;断裂;奥氏体; 渗碳体;断裂前;最大塑性变形;通用橡胶;特种橡胶;含碳量;万分之几;正火;碳钢;合金元素;酸性焊条;碱性焊条;表面淬火; 表面化学热处理;分离工序;整模造型;分模造型;变形工序;挖砂造型;活块造型;焊芯;药皮 题目1、 题目2强度就是指金属材料在外载荷作用下, 题目3 题目4在铁碳合金中, 题目5的最大应力值。 题目6优质碳素结构钢的牌号有两位数字表示, 题目7 题目8 题目9 题目10淬火前,若钢中存在网状渗碳体,,否则会增大钢的淬透性。 题目11 题目12根据药皮所含氧化物的性质, 题目13 题目14 二、就是非判断题(每题1分,共42分) 题目15冲击韧性值随温度的降低而增加。 选择一项:

对 错 题目16抗拉强度就是表示金属材料抵抗最大均匀塑性变形或断裂的能力。 选择一项: 对 错 题目17硬度就是指金属材料抵抗其她物体压入其表面的能力。 选择一项: 对 错 题目18金属材料在外载荷作用下产生断裂前所能承受最大塑性变形的能力称为塑性。选择一项: 对 错 题目19冲击韧性值随温度的降低而减小。 选择一项: 对 错 题目20强度越高,塑性变形抗力越大,硬度值也越高。 选择一项: 对 错 题目21屈服强度就是表示金属材料抵抗微量弹性变形的能力。 选择一项: 对 错 题目22冲击韧性值愈大,材料的韧性愈好。 选择一项: 对

错 题目23硬度就是指金属材料抵抗比它更硬的物体压入其表面的能力。选择一项: 对 错 题目24通常材料的力学性能就是选材的主要指标。 选择一项: 对 错 题目25一般来说,材料的硬度越高,耐磨性越好。 选择一项: 对 错 题目26测量布氏硬度时,压头为淬火钢球,用符号HBW表示。 选择一项: 对 错 题目27测量布氏硬度时,压头为淬火钢球,用符号HBS表示。 选择一项: 对 错 题目28测量布氏硬度时,压头为硬质合金球,用符号HBW表示。 选择一项: 对 错 题目29测量洛氏硬度时,压头为120°金刚石圆锥体,用符号HRC表示。选择一项: 对 错

国开《行政组织学》形考任务试题及答案

国开《行政组织学》形考任务试题及答案 形考任务一 一、判断题 1.在霍桑试验的基础上，梅奥于1933年出版了《工业文明中的人的问题》一书，系统地阐述了与古典管理理论截然不同的一些观点。\\对 2.阿吉里斯在《个性与组织》一书中提出了“不成熟-—成熟理论”。\\对 3.斯蒂格利茨由于在决策理论研究方面的贡献而荣获1978年诺贝尔经济学奖。\\错 4.马斯洛在其代表性著作《人类动机的理论》和《激励与个人》中，提出了著名的公平理论。\\错 5.美国学者巴纳德在1938年出版的《经理人员的职能》这本书中，系统地提出了动态平衡组织理论\对 6.社会系统组织理论的创始者为美国著名的社会学家罗森茨韦克。\\错 7.邓肯将组织环境分为内部环境和外部环境。\\对 8.卡斯特和罗森茨韦克将影响一切组织的一般环境特征划分为文化特征、技术特征、教育特征、政治特 征、法制特征、自然资源特征、人口特征、社会特征、经济特征等几个方面。\\对 9.组织界限以内与组织的个体决策行为直接相关的自然和社会因素被称为组织的内部环境。\\对 10.组织界限之外与组织内个体决策直接相关的自然和社会因素被称为组织的外部环境。\\对 11.“组织”一词，源自希腊文，1873年，哲学家斯宾塞用“组织”来指涉“已经组合的系统或社会”。\\对 12.与个别行政组织的决策转换过程相关联的更具体的力量被称为行政组织的工作环境。\\对 13.以明文规定的形式确立下来，成员具有正式分工关系的组织为非正式组织。\\错 14.以镇压、暴力等控制手段作为控制和管理下属的主要方式，此种类型的组织为规范性组织。\\错 15.以组织的参与者或成员为主要的受惠对象，组织的目的在于维护及促进组织成员所追求的利益，此种类型的组织为互利性组织。\\对 16.规范地讲，行政组织是追求行政权力的组织。\\错 17.韦伯是科学管理运动的先驱者，被誉为“科学管理之父”。\\错 18.1911年，泰勒发表了《科学管理原理》一书，掀起了一场科学管理的革命。\\对 19.行政管理学派的代表人物法约尔，被誉为“管理理论之父”。 \\对 20.德国著名的社会学家韦伯在《高级管理人员的职能》一书中，提出了理想型官僚组织理论。\\错

行政组织学形考任务5

形考任务五 一、判断题（正确划“√”，错误划“×”，每小题2分，共40分） 1.根据行政组织文化产生的时间，行政组织文化可以分为传统行政组织文化和当代行政组织文化。对 2.行政组织文化具有多种功能，它能把组织成员个人目标与组织目标有机结合起来，引导组织成员的行为，我们把这种功能称为控制功能。错 3.行政组织文化相比于正式的组织规章制度的控制作用，它具有软约束性的特性。对 4.行政组织文化是一种群体文化，是一种无形的管理方式。对 5.行政组织绩效就是指的行政组织活动的成果。错 6.经济性指标一般指行政组织投入到管理中的资源，其关心的是行政组织的投入。对 7.效果通常是指公共服务符合政策目标的程度，其关心的是手段。对 8.效率就是指投入与产出之间的比例，力求以最少的投入获得最大的产出，其关心的是手段问题。对 9.组织变革不是一个持续循环与发展的过程，因为要考虑到组织的稳定。对 10.组织发展起源于20世纪50年代初的调查反馈方法和实验室培训运动。它的先驱是法国心理学家烈文。对 11.1957年麦格雷戈应邀到联合碳化公司与公司人事部门联合成立顾问小组，把实验室训练的技术系统地在公司使用。这个小组后被称之

为“T训练小组”。错 12.作为一套极有系统的组织发展方案，格道式发展模式的目的在于使组织达到一种最佳状态。此模式创立者为布莱克和默顿。对 13.系统变革模式认为，组织是一个系统，是由技术、结构、人员和任务四个因素构成，任何一个因素的变化都会牵动和引起系统的变化。系统变革模式的创始人为利维特。对 14.美国心理学家埃德加·薛恩在其《组织心理学》一书中提出了系统变革模式。错 15.罗宾·斯特克兹认为，组织变革的方式取决于组织成员的技术能力和人际关系能力的组合，提出了渐进式变革模式。错 16.管理学大师德鲁克在《后资本主义社会》一书中指出：“世界上没有贫穷的国家，只有无知的国家”。对 17.知识经济与传统经济相比，知识成为组织根本的生产要素。对 18.组织理论家卡斯特和罗森茨韦克认为，未来的组织将更趋向于动态和灵活。对 19.战略管理的核心是问题管理。错 20.随着信息技术的发展，将信息科技运用于行政组织的管理，建立“节约型政府”已经成为各国的一个普遍趋势。错 二、多项选择题（每题有两个和两个以上正确答案，每小题3分，共60分） 21.根据其在行政组织中所占有的地位，行政组织文化可以分为___。主文化亚文化

江苏开放大学机械制造基础形考1

江苏开放大学 形成性考核作业学号 姓名 课程代码110036 课程名称机械制造基础评阅教师 第 1 次任务 共 5 次任务 江苏开放大学

一、选择题 1、金属材料在外力作用下抵抗塑性变形和断裂的能力叫（ B ）。 A．硬度 B．强度 C．塑性 D．弹性 2、高碳钢获得最佳切削性能的热处理工艺方法是（ B ）。 A.完全退火 B.球化退火 C.去应力退火 D.正火 3、08F钢中的平均含碳量为（ A ）。 A．0.08% B.0.8% C.8% 4、T10A钢按用途分属于（ C ）。 A．沸腾钢 B.优质钢 C.工具钢 5、在常用材料中，灰铸铁铸造性能（ A ）。 A．好 B．差 C．较差 6、无需填充金属，焊接变形小，生产率高的焊接方法是（ B ）。 A．埋弧焊 B．电阻焊 C．电弧焊 7、下列材料中（ B ）材料的焊接性能最差。 A．低碳钢 B．高碳钢 C．铸铁 8、机床床身的材料一般为（ B ）。 A．可锻铸铁 B.灰铸铁 C.球墨铸铁 D.蠕墨铸铁 9、下列金属属于轴承合金的是（ A ）。 A．铅青铜 B.硅青铜 C.铝镁合金 D.铝锌合金 10、生产柴油机曲轴应选材料是（ C ）。 A．HT200 B.QT700-2 C.KTH330-08 D.Q345 二、填空题 1、纯铁在加热过程中，其体心立方晶格与面心立方晶格的相互转变现象，称为同素异构转变。 2、一般情况下，晶粒越细，金属的强度、塑性和韧性越好。 3、奥氏体是碳在γ-Fe中的固溶体，呈面心立方晶格。 4、根据Fe-Fe3C 相图中成分-组织-性能的规律，选材时，建筑结构和各种型钢选塑性和韧性好的低碳钢，各种机械零件选用强度、塑性和韧性好的中碳钢，各种工具要用硬度高而耐磨性好的高碳钢。

《行政组织学》形考任务一答案

《行政组织学》形考任务一 一、判断题（正确划“√”，错误划“×”，每小题2分，共40分） 在霍桑试验的基础上，梅奥于1933年出版了《工业文明中的人的问题》一书，系统地阐述了与古典管理理论截然不同的一些观点。（√） 2. 阿吉里斯在《个性与组织》一书中提出了“不成熟-—成熟理论”。（√） 3.斯蒂格利茨由于在决策理论研究方面的贡献而荣获1978年诺贝尔经济学奖。（× ） 4.马斯洛在其代表性著作《人类动机的理论》和《激励与个人》中，提出了著名的公平理论。（× ） 5.美国学者巴纳德在1938年出版的《经理人员的职能》这本书中，系统地提出了动态平衡组织理论。（√） 6.社会系统组织理论的创始者为美国著名的社会学家罗森茨韦克。（× ） 7.邓肯将组织环境分为内部环境和外部环境。（√） 8.卡斯特和罗森茨韦克将影响一切组织的一般环境特征划分为文化特征、技术特征、教育特征、政治特征、法制特征、自然资源特征、人口特征、社会特征、经济特征等几个方面。（√） 9.组织界限以内与组织的个体决策行为直接相关的自然和社会因素被称为组织的内部环境。（√） 10.组织界限之外与组织内个体决策直接相关的自然和社会因素被称为组织的外部环境。（√） 11.“组织”一词，源自希腊文，1873年，哲学家斯宾塞用“组织”来指涉“已经组合的系统或社会”。（√）

12.与个别行政组织的决策转换过程相关联的更具体的力量被称为行政组织的工作环境。（√） 13.以明文规定的形式确立下来，成员具有正式分工关系的组织为非正式组织。（× ） 14.以镇压、暴力等控制手段作为控制和管理下属的主要方式，此种类型的组织为规范性组织。（× ） 15.以组织的参与者或成员为主要的受惠对象，组织的目的在于维护及促进组织成员所追求的利益，此种类型的组织为互利性组织。（√ ） 16.规范地讲，行政组织是追求行政权力的组织。（× ） 17.韦伯是科学管理运动的先驱者，被誉为“科学管理之父”。（× ） 18.1911年，泰勒发表了《科学管理原理》一书，掀起了一场科学管理的革命。（√ ） 19.行政管理学派的代表人物法约尔，被誉为“管理理论之父”。（√） 20.德国著名的社会学家韦伯在《高级管理人员的职能》一书中，提出了理想型官僚组织理论。（× ） 二、多项选择题（每题有两个和两个以上正确答案，每小题3分，共60分） 21.美国行为科学家赫茨伯格在其《工作的推力》和《工作与人性》等著作中，提出影响人的积极性的因素主要有______AC______。 A. 激励因素 B. 物质因素 C. 保健因素 D. 精神因素

国家开放大学《行政组织学》形考任务1试题

国家开放大学《行政组织学》形考任务1试题 在霍桑试验的基础上，梅奥于1933年出版了《工业文明中的人的问题》一书，系统地阐述了与古典管理理论截然不同的一些观点。（） 阿吉里斯在《个性与组织》一书中提出了不成熟 斯蒂格利茨由于在决策理论研究方面的贡献而荣获1978年诺贝尔经济学奖。（） 马斯洛在其代表性著作《人类动机的理论》和《激励与个人》中，提出了著名的公平理论。 美国学者巴纳德在1938年出版的《经理人员的职能》这本书中，系统地提出了动态平衡组织理论。（） 社会系统组织理论的创始者为美国著名的社会学家罗森茨韦克。（） 邓肯将组织环境分为内部环境和外部环境。（） 卡斯特和罗森茨韦克将影响一切组织的一般环境特征划分为文化特征、技术特征、教育特征、政治特征、法制特征、自然资源特征、人口特征、社会特征、经济特征等几个方面。（） 组织界限以内与组织的个体决策行为直接相关的自然和社会因素被称为组织的内部环境。 组织界限之外与组织内个体决策直接相关的自然和社会因素被称为组织的外部环境。（） 组织一词，源自希腊文，1873年，哲学家斯宾塞用组织来指涉已经组合的系统或社会。（） 与个别行政组织的决策转换过程相关联的更具体的力量被称为行政组织的工作环境。（） 以明文规定的形式确立下来，成员具有正式分工关系的组织为非正式组织。（） 以镇压、暴力等控制手段作为控制和管理下属的主要方式，此种类型的组织为规范性组织。 以组织的参与者或成员为主要的受惠对象，组织的目的在于维护及促进组织成员所追求的利益，此种类型的组织为互利性组织。（） 规范地讲，行政组织是追求行政权力的组织。（） 韦伯是科学管理运动的先驱者，被誉为科学管理之父。（） 1911年，泰勒发表了《科学管理原理》一书，掀起了一场科学管理的革命。（） 行政管理学派的代表人物法约尔，被誉为管理理论之父。（） 德国著名的社会学家韦伯在《高级管理人员的职能》一书中，提出了理想型官僚组织理论。 美国行为科学家赫茨伯格在其《工作的推力》和《工作与人性》等著作中，提出影响人的

机械制造基础形成性考核册作业1答案

机械制造基础形成性考核册作业1答案 1、常用的工程材料可以用教材第一页的表格表示，请完成工程材料的分类表： 答： 1、人们在描述金属材料力学性能重要指标时，经常使用如下术语，请填写其使用的符号和 内涵： （a）强度：金属材料在外载荷的作用下抵抗塑性变形和断裂的能力。强度有屈服强度σs和抗拉强度σb。 （b）塑性：金属材料在外载荷作用下产生断裂前所能承受最大塑性变形的能力。 （c）强度：是指金属材料抵抗比它更硬的物体压入其表面的能力。硬度有布氏硬度HBS和洛氏硬度HR。 （d）冲击韧性：金属抵抗冲击载荷作用而不被破坏的能力。 （e）疲劳强度：金属材料经受无数次交变载荷作用而不引起断裂的最大应力值。2、参照教材图1—2填写材料拉伸曲线中的相应特征值点的符号，并描述相应的含义。 （a）横坐标表示：试件的变形量ΔL （b）纵坐标表示：载荷F （c）S点：屈服点 （d）b点：颈缩点 （e）k点：拉断点 （f）屈服点以后出现的规律：试样的伸长率又随载荷的增加而增大，此时试样已产生较大的塑性变形，材料的抗拉强度明显增加。 3、一根标准试样的直径为10mm、标距长度为50mm。拉伸试验时测出试样在26kN时屈服， 出现的最大载荷为45 kN。拉断后的标距长度为58mm，断口处直径为7.75mm。试计算试样的σ0.2、σb。 答：σ0.2=F0.2/S0=26000/(3.14*0.0052)=3.3*108MPa σb=F b/S0=45000/(3.14*0.0052)=5.7*108MPa

4、HR是零件设计中常用的表示材料硬度指标。请回答下表中表示的有效范围和应用范围： HR和HB有什么差别？ 答：两种都是测试硬度的标准，区别在于测量方法不同。两种硬度标准根本性区别在于：布氏和洛氏测量的对象不同。布氏测量低硬度的材料，洛氏测量高硬度的材料。 6、参照教材1-7图补充如下Fe-Fe3C相图缺少的内容（标注相应的数据和材料相符号），思考每个线条和线条包容区域内金属相符号的特点 参考上图回答以下问题：： （a.）. AC线为：合金的液相线 （b.）. Acm线为：碳在奥氏体中的溶解限度线 （c.）. AC3线为：奥氏体中开始析出铁素体或铁素体全部溶入奥氏体的转变线 （d.）. A表示：奥氏体相 （e.）. F表示：铁素体相 （f.）. P表示：珠光体相 （g.）. Ld表示：液相 （h.）. Fe3C表示：渗碳体相 （i.）. 含碳0.77％为：都是奥氏体 （j.）. 导致过共析钢中材料脆性很高的主要原因是：若加热到略高于AC1温度时，珠光体完全转变成奥氏体，并又少量的渗碳体溶入奥氏体。此时奥氏体晶粒 细小，且其碳的质量分数已稍高与共析成分。如果继续升高温度，则二次渗

行政组织学形考任务四参考答案

一、判断题（正确划“√”，错误划“×”，每小题2分，共40分） 1、行政组织决策的目的是为了实现社会的共同利益。（√） 2、行政组织决策是以行政权力为后盾。√ 3、风险型决策的决策后果无法预测。× 4、确定目标是行政组织进行决策的起点。× 5、中枢系统是行政组织决策的中心。√ 6、美国政治学家伊斯顿被认为是决策理论研究的开创者。× 7、在决策理论研究领域，杜鲁门提出了团体决策模型。√ 8、现代观点认为，冲突既具有建设性又具有破坏性。√ 9、组织中最佳的冲突状态是没有冲突。× 10、解决冲突的基本策略中具有“治本”性的是正视策略。√ 11、合作意向都很高，宁可牺牲自身利益而使对方达到目的的冲突处理模式为协作型。× 12、通过组织明文规定的原则、渠道进行的信息传递和交流，是一种正式沟通。√ 13、信息的发讯者和受讯者以协商、会谈、讨论的方式进行信息的交流与意见反馈，直到双方共同了解为止，这种沟通形式为双向沟通。√ 14、组织系统中处于相同层次的人、群体、职能部门之间进行的信息传递和交流为平行沟通 √ 15、在组织管理中，书面沟通方式要优于口头沟通。× 16、作报告、发指示、下命令等属于单向沟通。√ 17、20世纪90年代初陈国权开始研究组织学习和学习型组织，并提出了组织学习系统理论（OLST）。√ 18、行政组织学习是一种全员学习。× 19、知识的主要构成要素包括经验、事实、判断以及经验法则。√ 20、行政组织学习不是组织内部成员个人学习的简单相加，而是一个社会过程。√ 二、多项选择题（每题有两个和两个以上正确答案，每小题3分，共60分） 21、根据决策所具有的条件的可靠程度的不同，决策可分为____。 确定型决策, 风险型决策, 不确定型决策 22、正确的决策目标应该具备的条件是___。定量化, 有一定的时间限制, 要明确责任 23、西蒙的决策过程包括___。 情报活动阶段, 设计活动阶段, 抉择活动阶段, 审查活动阶段 24、冲突的特性有_____。客观性, 主观性, 程度性 25、符合现代冲突观点的是____。 冲突本身没有好坏之分, 有些冲突对组织具有破坏性, 有些冲突对组织具有建设性 26、根据冲突发生的方向，可将冲突分为_________。 横向冲突, 纵向冲突, 直线/职能冲突 27、回避策略中，解决冲突的方法包括______. 忽视, 分离, 限制 28、减少冲突的策略主要有______。谈判, 设置超级目标, 第三方介入, 结构调整 29、从组织沟通的一般模式和组成要素来看，组织沟通具有以下几个特点：_____。 动态性, 互动性, 不可逆性, 环境制约性 30、以组织结构及其运行程序为依据和标准，组织信息沟通的形式和类型可划分为以下几种：______。下行沟通, 上行沟通, 平行沟通 31、根据沟通是否需要第三者中介传递，我们可将沟通划分为以下两种类型：_____。 直接沟通, 间接沟通

《行政组织学》形考任务1(参考答案)

《行政组织学》形考任务1（参考答案） 形考任务1 本次题型及题量如下： 一、判断题（共20道题，每题2分，共40分） 二、多项选择题（共20道题，每题3分，共60分） 全部做完后点击”提交所有答案并结束”，本次任务有三次答题机会，系统会记录最高分，请同学们珍惜机会，认真做题。允许试答次数：3；评分方法：最高分 一、判断题（正确划“√”，错误划“×”，每小题2分，共40分） 1、在霍桑试验的基础上，梅奥于1933年出版了《工业文明中的人的问题》一书，系统地阐述了与古典管理理论截然不同的一些观点。（）选择一项：√对，错 2、阿吉里斯在《个性与组织》一书中提出了“不成熟—成熟理论”。（）选择一项：√对，错 3、斯蒂格利茨由于在决策理论研究方面的贡献而荣获1978年诺贝尔经济学奖。（）选择一项：对，错 4、马斯洛在其代表性著作《人类动机的理论》和《激励与个人》中，提出了著名的公平理论。（）选择一项：对，错 5、美国学者巴纳德在1938年出版的《经理人员的职能》这本书中，系统地提出了动态平衡组织理论。（）选择一项：√对，错 6、社会系统组织理论的创始者为美国著名的社会学家罗森茨韦克。（）选择一项：对，错 7、邓肯将组织环境分为内部环境和外部环境。（）选择一项：√对，错 8、卡斯特和罗森茨韦克将影响一切组织的一般环境特征划分为文化特征、技术特征、教育特征、政治特征、法制特征、自然资源特征、人口特征、社会特征、经济特征等几个方面。（）选择一项：√对，错 9、组织界限以内与组织的个体决策行为直接相关的自然和社会因素被称为组织的内部环境。（）选择一项：√对，错 10、组织界限之外与组织内个体决策直接相关的自然和社会因素被称为组织的外部环境。（）选择一项：√对，错 11、“组织”一词，源自希腊文，1873年，哲学家斯宾塞用“组织”来指涉“已经组合的系统或社会”。（）选择一项：√对，错 12、与个别行政组织的决策转换过程相关联的更具体的力量被称为行政组织的工作环境。（）选择一项：√对，错 13、以明文规定的形式确立下来，成员具有正式分工关系的组织为非正式组织。（）选择一项：对，错 14、以镇压、暴力等控制手段作为控制和管理下属的主要方式，此种类型的组织为规范性组织。（）选择一项：对，错 15、以组织的参与者或成员为主要的受惠对象，组织的目的在于维护及促进组织成员所追求的利益，此种类型的组织为互利性组织。（）选择一项：√对，错 16、规范地讲，行政组织是追求行政权力的组织。（）选择一项：对，错 17、韦伯是科学管理运动的先驱者，被誉为“科学管理之父”。（）选择一项：对，错 18、1911年，泰勒发表了《科学管理原理》一书，掀起了一场科学管理的革命。（）

国家开放大学《机械制造基础》形考任务一试题

国家开放大学《机械制造基础》形考任务一试题 题目1：金属材料的力学性能是指在外载荷作用下其抵抗_____或_____的能力。 题目2：强度是指金属材料在外载荷作用下，抵抗_____和_____的能力。 题目3：金属材料在外载荷作用下产生_____所能承受_____的能力称为塑性。 题目4：在铁碳合金中，莱氏体是由_____和_____所构成的机械混合物。 题目5：疲劳强度是表示材料经受无数次_____作用而不引起_____的最大应力值。 题目6：优质碳素结构钢的牌号有两位数字表示，这两位数字具体表示钢中_____是_____。题目7：合金钢就是在_____的基础上有目的地加入一定量_____的钢。 题目8：橡胶按用途可分为_____和_____两大类。 题目9：常用的表面热处理工艺有_____和_____两种。 题目10：淬火前，若钢中存在网状渗碳体，应采用_____的方法予以消除，否则会增大钢的淬透性。 题目11：砂型铸造中常用的手工造型方有_____、_____、_____、_____等。 题目12：根据药皮所含氧化物的性质，焊条分为_____和_____两类。 题目13：冲压生产的基本工序有_____和_____两大类。 题目14：电焊条由_____和_____两部分组成。 题目15：冲击韧性值随温度的降低而增加。 题目16：抗拉强度是表示金属材料抵抗最大均匀塑性变形或断裂的能力。 题目17：硬度是指金属材料抵抗其他物体压入其表面的能力。 题目18：金属材料在外载荷作用下产生断裂前所能承受最大塑性变形的能力称为塑性。 题目19：冲击韧性值随温度的降低而减小。 题目20：强度越高，塑性变形抗力越大，硬度值也越高。 题目21：屈服强度是表示金属材料抵抗微量弹性变形的能力。 题目22：冲击韧性值愈大，材料的韧性愈好。 题目23：硬度是指金属材料抵抗比它更硬的物体压入其表面的能力。

行政组织学形考任务一参考答案

行政组织学形考任务一 一、判断题（正确划“√”，错误划“×”，每小题2分，共40分） 1、在霍桑试验的基础上，梅奥于1933年出版了《工业文明中的人的问题》一书，系 统地阐述了与古典管理理论截然不同的一些观点。（√） 2、阿吉里斯在《个性与组织》一书中提出了“不成熟—成熟理论”。（√） 3、斯蒂格利茨由于在决策理论研究方面的贡献而荣获1978年诺贝尔经济学奖。（×） 4、马斯洛在其代表性著作《人类动机的理论》和《激励与个人》中，提出了著名的公 平理论。（×） 5、美国学者巴纳德在1938年出版的《经理人员的职能》这本书中，系统地提出了动 态平衡组织理论。（√） 6、社会系统组织理论的创始者为美国著名的社会学家罗森茨韦克。（×） 7、邓肯将组织环境分为内部环境和外部环境。（√） 8、卡斯特和罗森茨韦克将影响一切组织的一般环境特征划分为文化特征、技术特征、 教育特征、政治特征、法制特征、自然资源特征、人口特征、社会特征、经济特征 等几个方面。（√） 9、组织界限以内与组织的个体决策行为直接相关的自然和社会因素被称为组织的内 部环境。（√） 10、组织界限之外与组织内个体决策直接相关的自然和社会因素被称为组织的外 部环境。（√） 11、“组织”一词，源自希腊文，1873年，哲学家斯宾塞用“组织”来指涉“已 经组合的系统或社会”。（√） 12、与个别行政组织的决策转换过程相关联的更具体的力量被称为行政组织的工

作环境。（√） 13、以明文规定的形式确立下来，成员具有正式分工关系的组织为非正式组织。 （×） 14、以镇压、暴力等控制手段作为控制和管理下属的主要方式，此种类型的组织为 规范性组织。（×） 15、以组织的参与者或成员为主要的受惠对象，组织的目的在于维护及促进组织成 员所追求的利益，此种类型的组织为互利性组织。（√） 16、规范地讲，行政组织是追求行政权力的组织。（×） 17、韦伯是科学管理运动的先驱者，被誉为“科学管理之父”。（×） 18、1911年，泰勒发表了《科学管理原理》一书，掀起了一场科学管理的革命。 （√） 19、行政管理学派的代表人物法约尔，被誉为“管理理论之父”。（√） 20、德国著名的社会学家韦伯在《高级管理人员的职能》一书中，提出了理想型官 僚组织理论。（×） 二、多项选择题 1、美国行为科学家赫茨伯格在其《工作的推力》和《工作与人性》等著作中，提出影响人 的积极性的因素主要有___激励因素, 保健因素____________。 2、西蒙指出，决策有两种极端的类型_______________。程序化决策, 非程序化决策 3、里格斯指出，“棱柱型社会”具有以下三个基本特征_________。异质性, 形式主义, 重 叠性

行政组织学形考任务一参考答案

行政组织学形考任务一 一、判断题(正确划“√”,错误划“×”,每小题2分,共40分) 1、在霍桑试验的基础上,梅奥于1933年出版了《工业文明中的人的问题》一书,系统地 阐述了与古典管理理论截然不同的一些观点。(√) 2、阿吉里斯在《个性与组织》一书中提出了“不成熟—成熟理论”。(√) 3、斯蒂格利茨由于在决策理论研究方面的贡献而荣获1978年诺贝尔经济学奖。(×) 4、马斯洛在其代表性著作《人类动机的理论》与《激励与个人》中,提出了著名的公平 理论。(×) 5、美国学者巴纳德在1938年出版的《经理人员的职能》这本书中,系统地提出了动态 平衡组织理论。(√) 6、社会系统组织理论的创始者为美国著名的社会学家罗森茨韦克。(×) 7、邓肯将组织环境分为内部环境与外部环境。(√) 8、卡斯特与罗森茨韦克将影响一切组织的一般环境特征划分为文化特征、技术特征、 教育特征、政治特征、法制特征、自然资源特征、人口特征、社会特征、经济特征等几个方面。(√) 9、组织界限以内与组织的个体决策行为直接相关的自然与社会因素被称为组织的内 部环境。(√) 10、组织界限之外与组织内个体决策直接相关的自然与社会因素被称为组织的外 部环境。(√) 11、“组织”一词,源自希腊文,1873年,哲学家斯宾塞用“组织”来指涉“已经组 合的系统或社会”。(√) 12、与个别行政组织的决策转换过程相关联的更具体的力量被称为行政组织的工

作环境。(√) 13、以明文规定的形式确立下来,成员具有正式分工关系的组织为非正式组织。 (×) 14、以镇压、暴力等控制手段作为控制与管理下属的主要方式,此种类型的组织为 规范性组织。(×) 15、以组织的参与者或成员为主要的受惠对象,组织的目的在于维护及促进组织成 员所追求的利益,此种类型的组织为互利性组织。(√) 16、规范地讲,行政组织就是追求行政权力的组织。(×) 17、韦伯就是科学管理运动的先驱者,被誉为“科学管理之父”。( ×) 18、1911年,泰勒发表了《科学管理原理》一书,掀起了一场科学管理的革命。(√) 19、行政管理学派的代表人物法约尔,被誉为“管理理论之父”。( √) 20、德国著名的社会学家韦伯在《高级管理人员的职能》一书中,提出了理想型官 僚组织理论。(×) 二、多项选择题 1、美国行为科学家赫茨伯格在其《工作的推力》与《工作与人性》等著作中,提出影响人 的积极性的因素主要有___激励因素, 保健因素____________。 2、西蒙指出,决策有两种极端的类型_______________。程序化决策, 非程序化决策 3、里格斯指出,“棱柱型社会”具有以下三个基本特征_________。异质性, 形式主义, 重叠 性 4、里格斯在她创立的“棱柱模式理论”中,将社会形态划分______________。农业社会, 棱 柱社会, 工业社会