Pub Date : 2003-10-15DOI: 10.1109/AIPR.2003.1284267
Neena Madan Davessar, Sunil Madan, Hardeep Singh
A new approach to segmentation of machine printed Gurmukhi text has been suggested. This approach can easily be extended to other Indian language scripts such as Devnagri and Bangla. Most of the characters in these scripts have horizontal lines at the top called headlines. Besides, there are cases in which the characters are found touching in the scanned image, just below the headline. To resolve these issues, a two-pass mechanism is used. In pass-one it approximates the segmentation point, while in pass-two the cutting point is optimized. This approach has been very successful in segmenting a pair as well as triplets of touching characters.
{"title":"A hybrid approach to character segmentation of Gurmukhi script characters","authors":"Neena Madan Davessar, Sunil Madan, Hardeep Singh","doi":"10.1109/AIPR.2003.1284267","DOIUrl":"https://doi.org/10.1109/AIPR.2003.1284267","url":null,"abstract":"A new approach to segmentation of machine printed Gurmukhi text has been suggested. This approach can easily be extended to other Indian language scripts such as Devnagri and Bangla. Most of the characters in these scripts have horizontal lines at the top called headlines. Besides, there are cases in which the characters are found touching in the scanned image, just below the headline. To resolve these issues, a two-pass mechanism is used. In pass-one it approximates the segmentation point, while in pass-two the cutting point is optimized. This approach has been very successful in segmenting a pair as well as triplets of touching characters.","PeriodicalId":176987,"journal":{"name":"32nd Applied Imagery Pattern Recognition Workshop, 2003. Proceedings.","volume":"17 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2003-10-15","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"124128711","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}
Pub Date : 2003-10-15DOI: 10.1109/AIPR.2003.1284272
Xiuwen Liu, Qiang Zhang
We present spectral histogram representations for visual modeling. Based on a generative process, the representation is derived by partitioning the frequency domain into small disjoint regions and assuming independence among the regions. This gives rise to a set of filters and a representation consisting of marginal distributions of those filter responses. A distinct advantage of our representation is that it can be effectively used for different classification and recognition tasks, which is demonstrated by experiments and comparisons in texture classification, face recognition, and appearance-based 3D object recognition. The marked improvement over existing methods justifies our principle that effective priori knowledge should be derived from physical generative processes.
{"title":"Spectral histogram representations for visual modeling","authors":"Xiuwen Liu, Qiang Zhang","doi":"10.1109/AIPR.2003.1284272","DOIUrl":"https://doi.org/10.1109/AIPR.2003.1284272","url":null,"abstract":"We present spectral histogram representations for visual modeling. Based on a generative process, the representation is derived by partitioning the frequency domain into small disjoint regions and assuming independence among the regions. This gives rise to a set of filters and a representation consisting of marginal distributions of those filter responses. A distinct advantage of our representation is that it can be effectively used for different classification and recognition tasks, which is demonstrated by experiments and comparisons in texture classification, face recognition, and appearance-based 3D object recognition. The marked improvement over existing methods justifies our principle that effective priori knowledge should be derived from physical generative processes.","PeriodicalId":176987,"journal":{"name":"32nd Applied Imagery Pattern Recognition Workshop, 2003. Proceedings.","volume":"5 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2003-10-15","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"131255548","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}
Pub Date : 2003-10-15DOI: 10.1109/AIPR.2003.1284244
Syed A. Rizvi, N. Nasrabadi
In this paper, we investigate several fusion techniques for designing a composite classifier to improve the performance (probability of correct classification) of FLIR ATR. In this research, we propose to use four ATR algorithms for fusion. The individual performance of the four contributing algorithms ranges from 73.5% to about 77% of probability of correct classification on the testing set. We propose to use Bayes classifier, committee of experts, stacked-generalization, winner-takes-all, and ranking-based fusion techniques for designing the composite classifiers. The experimental results show an improvement of more than 6.5% over the best individual performance.
{"title":"Fusion techniques for automatic target recognition","authors":"Syed A. Rizvi, N. Nasrabadi","doi":"10.1109/AIPR.2003.1284244","DOIUrl":"https://doi.org/10.1109/AIPR.2003.1284244","url":null,"abstract":"In this paper, we investigate several fusion techniques for designing a composite classifier to improve the performance (probability of correct classification) of FLIR ATR. In this research, we propose to use four ATR algorithms for fusion. The individual performance of the four contributing algorithms ranges from 73.5% to about 77% of probability of correct classification on the testing set. We propose to use Bayes classifier, committee of experts, stacked-generalization, winner-takes-all, and ranking-based fusion techniques for designing the composite classifiers. The experimental results show an improvement of more than 6.5% over the best individual performance.","PeriodicalId":176987,"journal":{"name":"32nd Applied Imagery Pattern Recognition Workshop, 2003. Proceedings.","volume":"22 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2003-10-15","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"114823351","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}
Pub Date : 2003-10-15DOI: 10.1109/AIPR.2003.1284254
A. Schaum
Conventional spatial imaging of the same object at different times or with different sensing modalities often requires the identification of corresponding points within a solid object. A mathematically similar problem occurs in the remote hyperspectral imaging of one scene at two widely separated time intervals. In both cases the information can be interpreted using linear vector spaces, and the differences in sensed signals can be modeled with linear transformations of these spaces. Here we explore first, how much can be deduced about the transformations based solely on the multivariate statistics of the two data sets. Then we solve application-specific models for each of conventional and hyperspectral applications.
{"title":"Data association for fusion in spatial and spectral imaging","authors":"A. Schaum","doi":"10.1109/AIPR.2003.1284254","DOIUrl":"https://doi.org/10.1109/AIPR.2003.1284254","url":null,"abstract":"Conventional spatial imaging of the same object at different times or with different sensing modalities often requires the identification of corresponding points within a solid object. A mathematically similar problem occurs in the remote hyperspectral imaging of one scene at two widely separated time intervals. In both cases the information can be interpreted using linear vector spaces, and the differences in sensed signals can be modeled with linear transformations of these spaces. Here we explore first, how much can be deduced about the transformations based solely on the multivariate statistics of the two data sets. Then we solve application-specific models for each of conventional and hyperspectral applications.","PeriodicalId":176987,"journal":{"name":"32nd Applied Imagery Pattern Recognition Workshop, 2003. Proceedings.","volume":"21 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2003-10-15","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"132875436","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}
Pub Date : 2003-10-15DOI: 10.1109/AIPR.2003.1284283
Mi-Ae Ko, Young-Mo Kim
This paper presents a simple and robust algorithm for vehicle's license plate recognition system. Based on template matching, this algorithm can be applied for real time recognition of license plates for vehicle surveillance system. The working principle is weight feature based hierarchical template evaluation. The performance of the proposed system has been evaluated on images acquired in real traffic conditions.
{"title":"License plate surveillance system using weighted template matching","authors":"Mi-Ae Ko, Young-Mo Kim","doi":"10.1109/AIPR.2003.1284283","DOIUrl":"https://doi.org/10.1109/AIPR.2003.1284283","url":null,"abstract":"This paper presents a simple and robust algorithm for vehicle's license plate recognition system. Based on template matching, this algorithm can be applied for real time recognition of license plates for vehicle surveillance system. The working principle is weight feature based hierarchical template evaluation. The performance of the proposed system has been evaluated on images acquired in real traffic conditions.","PeriodicalId":176987,"journal":{"name":"32nd Applied Imagery Pattern Recognition Workshop, 2003. Proceedings.","volume":"65 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2003-10-15","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"123321973","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}
Pub Date : 2003-10-15DOI: 10.1109/AIPR.2003.1284274
D. Hall
Increasingly, multi-sensor systems are being developed to collect, process, and disseminate image and non-image data. Applications include homeland security, monitoring of facilities, and military situation assessment. Fusion of image and non-image data has traditionally been performed with extensive human-in-the-loop involvement. Typically the image data are used as the "fundamental" data source with non-image data simply overlaid on the image data, or conversely the non-image data are treated as fundamental, and the image data are used to confirm the identity of observed entities. This paper discusses the problem of multi-sensor fusion and argues that new techniques are emerging that allows fusion of image and non-image data at multiple levels of inference from the "raw" data level, to the feature level, decision-level, and knowledge level.
{"title":"Perspectives on the fusion of image and non-image data","authors":"D. Hall","doi":"10.1109/AIPR.2003.1284274","DOIUrl":"https://doi.org/10.1109/AIPR.2003.1284274","url":null,"abstract":"Increasingly, multi-sensor systems are being developed to collect, process, and disseminate image and non-image data. Applications include homeland security, monitoring of facilities, and military situation assessment. Fusion of image and non-image data has traditionally been performed with extensive human-in-the-loop involvement. Typically the image data are used as the \"fundamental\" data source with non-image data simply overlaid on the image data, or conversely the non-image data are treated as fundamental, and the image data are used to confirm the identity of observed entities. This paper discusses the problem of multi-sensor fusion and argues that new techniques are emerging that allows fusion of image and non-image data at multiple levels of inference from the \"raw\" data level, to the feature level, decision-level, and knowledge level.","PeriodicalId":176987,"journal":{"name":"32nd Applied Imagery Pattern Recognition Workshop, 2003. Proceedings.","volume":"1 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2003-10-15","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"115857397","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}
Pub Date : 2003-10-15DOI: 10.1109/AIPR.2003.1284258
R. Easton, K. Knox, W. Christens-Barry
Multispectral imaging techniques are being applied to improve the readability of the text in a tenth-century manuscript that includes seven treatises of Archimedes. The manuscript was erased and overwritten about 200 years later with the text of a Christian prayer book. This talk reports on the results of the multispectral imaging techniques used on the Archimedes palimpsest.
{"title":"Multispectral imaging of the Archimedes palimpsest","authors":"R. Easton, K. Knox, W. Christens-Barry","doi":"10.1109/AIPR.2003.1284258","DOIUrl":"https://doi.org/10.1109/AIPR.2003.1284258","url":null,"abstract":"Multispectral imaging techniques are being applied to improve the readability of the text in a tenth-century manuscript that includes seven treatises of Archimedes. The manuscript was erased and overwritten about 200 years later with the text of a Christian prayer book. This talk reports on the results of the multispectral imaging techniques used on the Archimedes palimpsest.","PeriodicalId":176987,"journal":{"name":"32nd Applied Imagery Pattern Recognition Workshop, 2003. Proceedings.","volume":"49 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2003-10-15","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"121166189","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}
Pub Date : 2003-10-15DOI: 10.1109/AIPR.2003.1284269
Rajiv Kapoor, Aditya Dutta, D. Bagai, T. Kamal
The paper is a study demonstrating the application of discrete multiwavelets in medical image registration. The idea is to improve the image content by fusing images like MRI, CT and SPECT images, so as to provide more information to the doctor. The process of fusion is not new but here the results of study have been compared with the results from FCM algorithm used for similar application. Multiwavelets have been used for better clustering, as their decomposition results were better than Daubechies decomposition. A new feature based fusion algorithm has been used. This method shows results better than other methods for image registration when the images have been taken for the same person at a particular angle. The selective fusion not only gives more information but also helps in disease detection.
{"title":"Fusion for registration of medical images - a study","authors":"Rajiv Kapoor, Aditya Dutta, D. Bagai, T. Kamal","doi":"10.1109/AIPR.2003.1284269","DOIUrl":"https://doi.org/10.1109/AIPR.2003.1284269","url":null,"abstract":"The paper is a study demonstrating the application of discrete multiwavelets in medical image registration. The idea is to improve the image content by fusing images like MRI, CT and SPECT images, so as to provide more information to the doctor. The process of fusion is not new but here the results of study have been compared with the results from FCM algorithm used for similar application. Multiwavelets have been used for better clustering, as their decomposition results were better than Daubechies decomposition. A new feature based fusion algorithm has been used. This method shows results better than other methods for image registration when the images have been taken for the same person at a particular angle. The selective fusion not only gives more information but also helps in disease detection.","PeriodicalId":176987,"journal":{"name":"32nd Applied Imagery Pattern Recognition Workshop, 2003. Proceedings.","volume":"49 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2003-10-15","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"127929295","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}
Pub Date : 2003-10-15DOI: 10.1109/AIPR.2003.1284268
Li Tao, V. Asari
A luminance based multi scale retinex (LB/spl I.bar/MSR) algorithm for the enhancement of darker images is proposed in this paper. The new technique consists only the addition of the convolution results of 3 different scales. In this way, the color noise in the shadow/dark areas can be suppressed and the convolutions with different scales can be calculated simultaneously to save CPU time. Color saturation adjustment for producing more natural colors is implemented. Each spectral band can be adjusted based on the enhancement of the intensity of the band and by using a color saturation parameter. The color saturation degree can be automatically adjusted according to different types of images by compensating the original color saturation in each band. Luminance control is applied to prevent the unwanted luminance drop at the uniform luminance areas by automatically detecting the luminance drop and keeping the luminance up to certain level that is evaluated from the original image. Down-sized convolution is used for fast processing and then the result is re-sized back to the original size. Performance of the new enhancement algorithm is tested in various images captured at different lighting conditions. It is observed that the new technique outperforms the conventional MSR technique in terms of the quality of the enhanced images and computational speed.
{"title":"Modified luminance based MSR for fast and efficient image enhancement","authors":"Li Tao, V. Asari","doi":"10.1109/AIPR.2003.1284268","DOIUrl":"https://doi.org/10.1109/AIPR.2003.1284268","url":null,"abstract":"A luminance based multi scale retinex (LB/spl I.bar/MSR) algorithm for the enhancement of darker images is proposed in this paper. The new technique consists only the addition of the convolution results of 3 different scales. In this way, the color noise in the shadow/dark areas can be suppressed and the convolutions with different scales can be calculated simultaneously to save CPU time. Color saturation adjustment for producing more natural colors is implemented. Each spectral band can be adjusted based on the enhancement of the intensity of the band and by using a color saturation parameter. The color saturation degree can be automatically adjusted according to different types of images by compensating the original color saturation in each band. Luminance control is applied to prevent the unwanted luminance drop at the uniform luminance areas by automatically detecting the luminance drop and keeping the luminance up to certain level that is evaluated from the original image. Down-sized convolution is used for fast processing and then the result is re-sized back to the original size. Performance of the new enhancement algorithm is tested in various images captured at different lighting conditions. It is observed that the new technique outperforms the conventional MSR technique in terms of the quality of the enhanced images and computational speed.","PeriodicalId":176987,"journal":{"name":"32nd Applied Imagery Pattern Recognition Workshop, 2003. Proceedings.","volume":"27 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2003-10-15","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"125245997","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}
Pub Date : 2003-10-15DOI: 10.1109/AIPR.2003.1284243
N. Gupta
We report on the development and characterization of a compact, lightweight, robust, and field-portable spectropolarimetric imaging system at the U.S. Army Research Laboratory (ARL). It operates in the 400 to 900 nm region with a passband of 10 nm at 600 nm. This automated imager is designed using a tellurium dioxide (TeO/sub 2/) acousto-optic tunable filter (AOTF) as an agile spectral selection element and a commercial nematic liquid-crystal variable retardation (LCVR) plate as a tunable polarization selection device with an off-the-shelf uncooled charge coupled device (CCD) camera and optics. Image acquisition with both spectral and polarization features facilitates significant improvement in target detection. This paper has described the design concept and a program, with a detailed description of the VNIR imager, and present images obtained from it and the analysis of the results.
{"title":"Fused spectropolarimetric visible near-IR imaging","authors":"N. Gupta","doi":"10.1109/AIPR.2003.1284243","DOIUrl":"https://doi.org/10.1109/AIPR.2003.1284243","url":null,"abstract":"We report on the development and characterization of a compact, lightweight, robust, and field-portable spectropolarimetric imaging system at the U.S. Army Research Laboratory (ARL). It operates in the 400 to 900 nm region with a passband of 10 nm at 600 nm. This automated imager is designed using a tellurium dioxide (TeO/sub 2/) acousto-optic tunable filter (AOTF) as an agile spectral selection element and a commercial nematic liquid-crystal variable retardation (LCVR) plate as a tunable polarization selection device with an off-the-shelf uncooled charge coupled device (CCD) camera and optics. Image acquisition with both spectral and polarization features facilitates significant improvement in target detection. This paper has described the design concept and a program, with a detailed description of the VNIR imager, and present images obtained from it and the analysis of the results.","PeriodicalId":176987,"journal":{"name":"32nd Applied Imagery Pattern Recognition Workshop, 2003. Proceedings.","volume":"167 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2003-10-15","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"116331457","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：