Vision systems typically incorporate object rigidity constraints. Such constraints are meaningful in manufacturing processes, but not in the natural visual world, where elastic objects, jointed objects, a variety of viewing geometry deformations, and variations of structure within an object class are common. This work extends the capabilities of recognition algorithms to the realm of flexible objects, discussing the development of algorithms, data structures and mathematical models for the recognition of flexible objects from a single image.<>
{"title":"Recognition of flexible objects","authors":"R. Carlson","doi":"10.1109/IAI.1994.336677","DOIUrl":"https://doi.org/10.1109/IAI.1994.336677","url":null,"abstract":"Vision systems typically incorporate object rigidity constraints. Such constraints are meaningful in manufacturing processes, but not in the natural visual world, where elastic objects, jointed objects, a variety of viewing geometry deformations, and variations of structure within an object class are common. This work extends the capabilities of recognition algorithms to the realm of flexible objects, discussing the development of algorithms, data structures and mathematical models for the recognition of flexible objects from a single image.<<ETX>>","PeriodicalId":438137,"journal":{"name":"Proceedings of the IEEE Southwest Symposium on Image Analysis and Interpretation","volume":"103 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"1994-04-21","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"124159774","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
A method to extract shape features based on corners is described. Corners contain most of the shape information. Extraction of shape features which are invariant to scaling, rotation and translation is an important problem in computer vision and automatic target recognition systems. A Canny (1986) edge detector which is capable of producing single pixel wide edges is used for obtaining the contour from an image. Using this closed contour as input, the arch height function is computed at each point. The local maxima's correspond to the corner points in the shape. A set of efficient one dimensional moments which are invariant under rotation, translation and scale change is computed. These are the corresponding shape features. Classification is achieved by comparing the extracted features with the shape feature library. In order to validate the concept the following experiments were performed. Ten dissimilar aircrafts and ten similar aircrafts were used as inputs. Contour based moments performed better than the geometric moments in both the data sets. Rotation invariance of two very similar aircrafts showed that contour based moments performed better. The procedure described provides an elegant approach for extracting shape features. These features can also be used as inputs for training and recognizing shapes using neural networks.<>
{"title":"Shape feature extraction from object corners","authors":"K. K. Rao, R. Krishnan","doi":"10.1109/IAI.1994.336665","DOIUrl":"https://doi.org/10.1109/IAI.1994.336665","url":null,"abstract":"A method to extract shape features based on corners is described. Corners contain most of the shape information. Extraction of shape features which are invariant to scaling, rotation and translation is an important problem in computer vision and automatic target recognition systems. A Canny (1986) edge detector which is capable of producing single pixel wide edges is used for obtaining the contour from an image. Using this closed contour as input, the arch height function is computed at each point. The local maxima's correspond to the corner points in the shape. A set of efficient one dimensional moments which are invariant under rotation, translation and scale change is computed. These are the corresponding shape features. Classification is achieved by comparing the extracted features with the shape feature library. In order to validate the concept the following experiments were performed. Ten dissimilar aircrafts and ten similar aircrafts were used as inputs. Contour based moments performed better than the geometric moments in both the data sets. Rotation invariance of two very similar aircrafts showed that contour based moments performed better. The procedure described provides an elegant approach for extracting shape features. These features can also be used as inputs for training and recognizing shapes using neural networks.<<ETX>>","PeriodicalId":438137,"journal":{"name":"Proceedings of the IEEE Southwest Symposium on Image Analysis and Interpretation","volume":"68 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"1994-04-21","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"121437390","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
The authors propose a new method for calibrating the head-to-eye relation of a binocular head with very high accuracy. This method consists of a closed-form solution for initial estimation and an iteration procedure for accuracy refinement. The closed-form solution is already quite accurate (more accurate than the existing method) when the number of calibration input data is large enough or when the amount of measurement noise is small. Using the closed-form solution as an initial estimate, the iterative method can be adopted to further improve the calibration accuracy. Instead of using the estimates of some transformation matrices as the calibration input data, which is needed in almost all the existing techniques, the proposed method needs only to measure the 3D coordinates of a single calibration point. This avoids the difficulty in finding a reliable extrinsic camera calibration technique. Also, during the calibration, the binocular head can be moved in a larger range while keeping the single calibration point in the viewfield of the camera, which in turn provides richer information for the head/eye calibration. According to their experiments, the proposed method achieved much higher accuracy than the existing techniques.<>
{"title":"Head/eye calibration of a binocular head by use of single calibration point","authors":"S. Shih, Y. Hung, Wei-Song Lin","doi":"10.1109/IAI.1994.336666","DOIUrl":"https://doi.org/10.1109/IAI.1994.336666","url":null,"abstract":"The authors propose a new method for calibrating the head-to-eye relation of a binocular head with very high accuracy. This method consists of a closed-form solution for initial estimation and an iteration procedure for accuracy refinement. The closed-form solution is already quite accurate (more accurate than the existing method) when the number of calibration input data is large enough or when the amount of measurement noise is small. Using the closed-form solution as an initial estimate, the iterative method can be adopted to further improve the calibration accuracy. Instead of using the estimates of some transformation matrices as the calibration input data, which is needed in almost all the existing techniques, the proposed method needs only to measure the 3D coordinates of a single calibration point. This avoids the difficulty in finding a reliable extrinsic camera calibration technique. Also, during the calibration, the binocular head can be moved in a larger range while keeping the single calibration point in the viewfield of the camera, which in turn provides richer information for the head/eye calibration. According to their experiments, the proposed method achieved much higher accuracy than the existing techniques.<<ETX>>","PeriodicalId":438137,"journal":{"name":"Proceedings of the IEEE Southwest Symposium on Image Analysis and Interpretation","volume":"1 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"1994-04-21","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"130732529","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Gibbs distributions have been widely used in the Bayesian approach to many image processing problems. However, little attention has been paid to whether or not the Gibbs distribution indeed models the images that occur in the particular area of application. Indeed, random samples from many of the proposed Gibbs distributions are likely to be uniformly smooth, and thus atypical for any application area. The authors investigate the possibility of finding Gibbs distributions which truly model certain global properties of images. Specifically, they construct a Gibbs distribution which models an image that consist of piecewise homogeneous regions by including different orders of neighbor interactions. By sampling the Gibbs distribution which arises from the model, they obtain images with piecewise homogeneous regions resembling the global features of the image that they intend to model; hence such a Gibbs distribution is indeed "image-modeling". They assess the adequacy of their model using a /spl chisup 2/ goodness-of-fit test. They also address how parameters are selected based on given image data. Importantly, the most essential parameter of the image model (related to the regularization parameter) is estimated in the process of constructing the image model. Comparative results are presented of the outcome of using their model and an alternative model as the prior in some image restoration problems in which noisy synthetic images were considered.<>
{"title":"Image-modeling Gibbs distributions for Bayesian restoration","authors":"M. Chan, E. Levitan, G. Herman","doi":"10.1109/IAI.1994.336691","DOIUrl":"https://doi.org/10.1109/IAI.1994.336691","url":null,"abstract":"Gibbs distributions have been widely used in the Bayesian approach to many image processing problems. However, little attention has been paid to whether or not the Gibbs distribution indeed models the images that occur in the particular area of application. Indeed, random samples from many of the proposed Gibbs distributions are likely to be uniformly smooth, and thus atypical for any application area. The authors investigate the possibility of finding Gibbs distributions which truly model certain global properties of images. Specifically, they construct a Gibbs distribution which models an image that consist of piecewise homogeneous regions by including different orders of neighbor interactions. By sampling the Gibbs distribution which arises from the model, they obtain images with piecewise homogeneous regions resembling the global features of the image that they intend to model; hence such a Gibbs distribution is indeed \"image-modeling\". They assess the adequacy of their model using a /spl chisup 2/ goodness-of-fit test. They also address how parameters are selected based on given image data. Importantly, the most essential parameter of the image model (related to the regularization parameter) is estimated in the process of constructing the image model. Comparative results are presented of the outcome of using their model and an alternative model as the prior in some image restoration problems in which noisy synthetic images were considered.<<ETX>>","PeriodicalId":438137,"journal":{"name":"Proceedings of the IEEE Southwest Symposium on Image Analysis and Interpretation","volume":"31 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"1994-04-21","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"117106042","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
The analysis of video images in stereo can extend machine vision to interpret the 3-D structure of a scene. Applications of stereo vision include robotics, industrial automation, autonomous land rovers and automated cartography. The simplest stereo paradigm, binocular stereo vision, provides man and many animals the capability to see the depth from two images without ambiguity. Thus, it is interesting to study the biological solution to stereopsis. In this paper, a biologically motivated model of stereopsis based on a coarse-to-fine matching algorithm using multiband Gabor wavelets is presented. This approach generates a dense disparity map by phase difference computation between stereo image pairs without complex feature extraction. Results of the algorithm for both synthetic and natural stereo images are presented.<>
{"title":"Stereo vision using Gabor wavelets","authors":"Tieh-Yuh Chen, W.N. Klarquist, A. Bovik","doi":"10.1109/IAI.1994.336690","DOIUrl":"https://doi.org/10.1109/IAI.1994.336690","url":null,"abstract":"The analysis of video images in stereo can extend machine vision to interpret the 3-D structure of a scene. Applications of stereo vision include robotics, industrial automation, autonomous land rovers and automated cartography. The simplest stereo paradigm, binocular stereo vision, provides man and many animals the capability to see the depth from two images without ambiguity. Thus, it is interesting to study the biological solution to stereopsis. In this paper, a biologically motivated model of stereopsis based on a coarse-to-fine matching algorithm using multiband Gabor wavelets is presented. This approach generates a dense disparity map by phase difference computation between stereo image pairs without complex feature extraction. Results of the algorithm for both synthetic and natural stereo images are presented.<<ETX>>","PeriodicalId":438137,"journal":{"name":"Proceedings of the IEEE Southwest Symposium on Image Analysis and Interpretation","volume":"1 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"1994-04-21","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"128982472","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
The authors present techniques for improving the contrast in spectrally variant images of mailpieces. These techniques were investigated for a system that captures mailpiece images and performs address interpretation in real-time. Hence it was necessary to satisfy both time constraint and an enhancement criterion. The methods developed are fast and are suitable for real-time applications. The large spectral variation of mailpiece images results in images of low-contrast. Examples are images of dark-colored envelopes with dark ink, such as greeting card envelopes, specially-designed envelopes and so on. This necessitates contrast enhancement methods to be highly adaptive. In the algorithms described most of the parameters are estimated dynamically from image characteristics. Conventional contrast enhancement methods are designed for a static environment and not amenable for use within this problem domain. Their method improves on previous work described by DuVall (1979) in compensating for the limited spectral response of the imaging system.<>
{"title":"Contrast enhancement for mailpiece images","authors":"G. Srikantan, R. Fenrich, S. Srihari","doi":"10.1109/IAI.1994.336680","DOIUrl":"https://doi.org/10.1109/IAI.1994.336680","url":null,"abstract":"The authors present techniques for improving the contrast in spectrally variant images of mailpieces. These techniques were investigated for a system that captures mailpiece images and performs address interpretation in real-time. Hence it was necessary to satisfy both time constraint and an enhancement criterion. The methods developed are fast and are suitable for real-time applications. The large spectral variation of mailpiece images results in images of low-contrast. Examples are images of dark-colored envelopes with dark ink, such as greeting card envelopes, specially-designed envelopes and so on. This necessitates contrast enhancement methods to be highly adaptive. In the algorithms described most of the parameters are estimated dynamically from image characteristics. Conventional contrast enhancement methods are designed for a static environment and not amenable for use within this problem domain. Their method improves on previous work described by DuVall (1979) in compensating for the limited spectral response of the imaging system.<<ETX>>","PeriodicalId":438137,"journal":{"name":"Proceedings of the IEEE Southwest Symposium on Image Analysis and Interpretation","volume":"15 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"1994-04-21","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"133199405","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
The problem of recognition of handwritten segmented digits irrespective of their size or stroke width is considered. A new approach of combining several different multi-layer perceptron (MLP) neural network classifiers operating on the same image is developed. The classification decisions made by individual MLPs are combined through a method called "behavior-knowledge space" (BKS). The BKS method relies on the behavior of the classifiers on the training set. The pseudo-Zernike moments extracted from the normalized and thinned image of the digit within its bounding circle are used as features. The approach is tested on 3000 digits using three classifiers and a hard error rate of 1.37% is obtained. This is a reduction of almost 50% compared to a single MLP network classifier. The results are also compared to an alternative method of combining the classifiers.<>
{"title":"Hand written digit recognition using BKS combination of neural network classifiers","authors":"A. Khotanzad, C. Chung","doi":"10.1109/IAI.1994.336676","DOIUrl":"https://doi.org/10.1109/IAI.1994.336676","url":null,"abstract":"The problem of recognition of handwritten segmented digits irrespective of their size or stroke width is considered. A new approach of combining several different multi-layer perceptron (MLP) neural network classifiers operating on the same image is developed. The classification decisions made by individual MLPs are combined through a method called \"behavior-knowledge space\" (BKS). The BKS method relies on the behavior of the classifiers on the training set. The pseudo-Zernike moments extracted from the normalized and thinned image of the digit within its bounding circle are used as features. The approach is tested on 3000 digits using three classifiers and a hard error rate of 1.37% is obtained. This is a reduction of almost 50% compared to a single MLP network classifier. The results are also compared to an alternative method of combining the classifiers.<<ETX>>","PeriodicalId":438137,"journal":{"name":"Proceedings of the IEEE Southwest Symposium on Image Analysis and Interpretation","volume":"33 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"1994-04-21","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"127133250","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
A maximum likelihood estimation (MLE) method is used to estimate the fractal dimension of a number of natural texture images with and without the presence of noise. An additional texture measure which can be linked to the lacunarity measure is used to characterize natural textures since fractal dimension alone cannot totally characterize texture images. Segmentation of natural textures is successfully achieved by a k-means clustering algorithm using fractal dimension and the additional measure as representative features.<>
{"title":"Analysis of texture images using robust fractal description","authors":"N. Avadhanam, S. Mitra","doi":"10.1109/IAI.1994.336692","DOIUrl":"https://doi.org/10.1109/IAI.1994.336692","url":null,"abstract":"A maximum likelihood estimation (MLE) method is used to estimate the fractal dimension of a number of natural texture images with and without the presence of noise. An additional texture measure which can be linked to the lacunarity measure is used to characterize natural textures since fractal dimension alone cannot totally characterize texture images. Segmentation of natural textures is successfully achieved by a k-means clustering algorithm using fractal dimension and the additional measure as representative features.<<ETX>>","PeriodicalId":438137,"journal":{"name":"Proceedings of the IEEE Southwest Symposium on Image Analysis and Interpretation","volume":"66 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"1994-04-21","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"130694279","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
The authors present a nonlinear algorithm for critical point detection (CPD). The algorithm eliminates the problems arising from curvature approximation and Gaussian filtering in the existing algorithms. By defining a "critical level" as the modified area confined by three consecutive "pseudo-critical points", a simple but very effective algorithm is developed. The comparison of the experimental results with those of many other CPD algorithms shows that the proposed algorithm is superior in all tested contours.<>
{"title":"A nonlinear algorithm for critical point detection","authors":"Peng Fei Zhu, P. Chirlian","doi":"10.1109/IAI.1994.336687","DOIUrl":"https://doi.org/10.1109/IAI.1994.336687","url":null,"abstract":"The authors present a nonlinear algorithm for critical point detection (CPD). The algorithm eliminates the problems arising from curvature approximation and Gaussian filtering in the existing algorithms. By defining a \"critical level\" as the modified area confined by three consecutive \"pseudo-critical points\", a simple but very effective algorithm is developed. The comparison of the experimental results with those of many other CPD algorithms shows that the proposed algorithm is superior in all tested contours.<<ETX>>","PeriodicalId":438137,"journal":{"name":"Proceedings of the IEEE Southwest Symposium on Image Analysis and Interpretation","volume":"1 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"1994-04-21","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"130371579","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
The authors briefly describe their new morphological operators of "boundary erosion region dilation (BERD)" and "boundary dilation region erosion (BDRE)". The paper covers definitions, properties, and some of the applications of these operators. BERD and BDRE operations may be used to approximately perform the job of various common morphological filtering operations with lesser computational times. BERD and BDRE operations with 2-D structuring elements may be used to construct robust connectivity preserving filters. Unlike commonly used median closing and opening operators/sup 1,5,6/, these filters do not destroy thin regions which rapidly change their direction or the regions which can not be probed by straight lines in limited directions. They have used these connectivity preserving filters for removing speckle noise from SAR images while retaining thin regions and fine details of regions boundaries.<>
{"title":"New morphological operators: BERD and BDRE","authors":"A. Kher, S. Mitra","doi":"10.1109/IAI.1994.336674","DOIUrl":"https://doi.org/10.1109/IAI.1994.336674","url":null,"abstract":"The authors briefly describe their new morphological operators of \"boundary erosion region dilation (BERD)\" and \"boundary dilation region erosion (BDRE)\". The paper covers definitions, properties, and some of the applications of these operators. BERD and BDRE operations may be used to approximately perform the job of various common morphological filtering operations with lesser computational times. BERD and BDRE operations with 2-D structuring elements may be used to construct robust connectivity preserving filters. Unlike commonly used median closing and opening operators/sup 1,5,6/, these filters do not destroy thin regions which rapidly change their direction or the regions which can not be probed by straight lines in limited directions. They have used these connectivity preserving filters for removing speckle noise from SAR images while retaining thin regions and fine details of regions boundaries.<<ETX>>","PeriodicalId":438137,"journal":{"name":"Proceedings of the IEEE Southwest Symposium on Image Analysis and Interpretation","volume":"70 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"1994-04-21","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"116561817","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
京公网安备 11010802042870号
京ICP备2023020795号-1
扫码关注我们
求助内容：
应助结果提醒方式：