A new distance measure is introduced for comparing polygonal curves which may be geometrically distorted relative to one another. The measure includes both directional and positional differences between segments and i...
详细信息
A new distance measure is introduced for comparing polygonal curves which may be geometrically distorted relative to one another. The measure includes both directional and positional differences between segments and is relatively insensitive to segmentation irregularity. Segment pairing is implemented with a conventional dynamic programming technique. Polygonal approximation allows great search reduction over pixel-based matching, thus making the technique practical for use on small computers. In addition, the measure is compatible with standard pruning techniques for even greater increases in speed. Some preliminary results in regard to handprinted character matching are shown.
Shape description and recognition is an important and interesting problem in scene analysis. The authors approach to shape description is a formal model of a shape consisting of a set of primitives, their properties, ...
详细信息
Shape description and recognition is an important and interesting problem in scene analysis. The authors approach to shape description is a formal model of a shape consisting of a set of primitives, their properties, and their interrelationships. The primitives are the simple parts and intrusions of the shape which can be derived through the graph-theoretic clustering procedure previously described. The interrelationships are two ternary relations on the primitives: the intrusion relation which relates two simple parts that join to the intrusion they surround and the protrusion relation which relates two intrusions to the protrusion between them. Using this model, a shape matching procedure that uses a tree search with look-ahead to find mappings from a prototype shape to a candidate shape has been developed.
In routine examination of chest radiographs, radiologists miss about 30% of the lung tumors in these pictures. To aid the radiologist in detecting these tumors the authors have designed and constructed, and are contin...
详细信息
In routine examination of chest radiographs, radiologists miss about 30% of the lung tumors in these pictures. To aid the radiologist in detecting these tumors the authors have designed and constructed, and are continuing to develop, a system consisting of a scanning microdensitometer, a digital minicomputer, a computer program, and a digital cathode-ray-tube color display. This system detects nodules in the range 5mm to 25mm mm and classifies them into two classes: tumor and nontumor. On a data base of 14 nodules in five radiographs, our system missed only one nodule and produced an average of seven false positives per radiograph.
In order to carry out an automatic procedure to analyze a sequence of images and extract in that way the dynamic characteristics of the motion represented in the sequence, a certain number of methods are, at present, ...
详细信息
In order to carry out an automatic procedure to analyze a sequence of images and extract in that way the dynamic characteristics of the motion represented in the sequence, a certain number of methods are, at present, considered and evaluated by many authors. In that context one of the more interesting aspects consists in obtaining the segmentation of each frame that could be considered ″normalized″ in a given interval of frames belonging to the sequence. A method to obtain a normalized segmentation for the special case considered, is presented. The technique is based on the iterated application of a pair of algorithms the first of which is able to identify the moving parts and to make a check on the rightness of such identification;the second one carries out the ″normalization″ of frames of the sequence on the basis of the results furnished by the first algorithm.
The k-syntactic similarity approach is couched in graphical representation terms and its ability to provide global recognition capability while retaining a low time complexity is explored. One potential application do...
详细信息
The k-syntactic similarity approach is couched in graphical representation terms and its ability to provide global recognition capability while retaining a low time complexity is explored. One potential application domain, that of composite shape decomposition into approximately convex subshapes, is described. This is shown to be equivalent to finding cycles within a particular graph. The approach yields valid decompositions in many cases where additional semantic considerations are necessary for proper analysis. The permissible graph structures representing composite shapes given a reasonable set of relations are determined. Experimental results on non-ideal data are given.
The problem of estimating the error probability of a given classification system is considered. Statistical properties of the empirical error count (C) and the average conditional error (R) estimators are studied. It ...
详细信息
The problem of estimating the error probability of a given classification system is considered. Statistical properties of the empirical error count (C) and the average conditional error (R) estimators are studied. It is shown that in the large sample case the R estimator is unbiased and its variance is less than that of the C estimator. In contrast to conventional methods of Bayes error estimation the unbiasedness of the R estimator for a given classifier can be obtained only at the price of an additional set of classified samples. On small test sets the R estimator may be subject to a pessimistic bias caused by the averaging phenomenon characterising the functioning of conditional error estimators.
Large images are becoming more and more common in earth resources monitoring, medical diagnosis and other applications. Often it would be helpful to work with only a subset of a large image since less space and time w...
详细信息
Large images are becoming more and more common in earth resources monitoring, medical diagnosis and other applications. Often it would be helpful to work with only a subset of a large image since less space and time would be required to process it. Subsets extracted according to semantic attributes have irregular shapes and as such are awkward to store and process. Irregular subsets can be covered with rectangular regions to simplify the regions to be stored and processed. Then the rectangular regions must be organized with an index. Here several covering methods are compared and indexing methods suggested. A surprizing result is that the sequential-greatest coverage heuristic can lead to arbitrarily bad coverings in some situations. However, this disadvantage can be overcome by combination with a tiling approach.
Future generation x-ray computed tomography scanners will be characterized by their ability to record simultaneously a sufficient number of x-ray projections to allow reconstructions of multiple adjacent cross section...
详细信息
Future generation x-ray computed tomography scanners will be characterized by their ability to record simultaneously a sufficient number of x-ray projections to allow reconstructions of multiple adjacent cross sections of the object under study. An ability to repeat the entire data collection procedure with great rapidity, allowing many scan passes per second, should encourage research and diagnostic studies of moving organs such as the heart and lungs in truly three dimensions and in real time. A combined series of algorithmic, software, special-purpose computer architecture, and hardware implementation studies have demonstrated significant progress toward computed tomography reconstruction processing.rates of 10**9 to 10**1**0 arithmetic operations per second.
Sequences of pictures of growing colonies of animal cells or microorganisms contain information useful to biologists. Special lighting increases the information contained in the pictures and decreases the difficulty o...
详细信息
Sequences of pictures of growing colonies of animal cells or microorganisms contain information useful to biologists. Special lighting increases the information contained in the pictures and decreases the difficulty of the image analysis. A high resolution flying spot scanner digitizes images of petri dishes recorded on film. Nonlinear edge region detectors and a region growing algorithm find the outlines of the colonies. Outlines which pass tests for validity are analyzed to extract further information. Data files of parameters are created which are analyzed by programs written by biologists. Statistical and other techniques were used to classify species and locate mutant colonies.
The widely used stepwise discriminant analysis procedure selects one variable at a time. However, instances arise in which variables occur naturally paired, such as the real and imaginary parts of a Fourier transforme...
详细信息
The widely used stepwise discriminant analysis procedure selects one variable at a time. However, instances arise in which variables occur naturally paired, such as the real and imaginary parts of a Fourier transformed signal. The existing stepwise discriminant procedure might select the real part of the ith harmonic and the imaginary part of the jth harmonic. This paper presents an algorithm that selects variables in pairs. With this algorithm, both the phase and magnitude of a signal can be used to form a discriminant function, as compared to the existing method of using the power only. From the simulation study, the discriminant procedure is better for inherently paired variables.
暂无评论