Pub Date : 2015-11-01DOI: 10.1109/SOCPAR.2015.7492767
J. Sato, T. Akashi
Nowadays, many computer vision techniques are applied to practical applications, such as surveillance and facial recognition systems. Some of such applications focus on information extraction from the human beings. However, people may feel psychological stress about recording their personal information, such as a face, behavior, and cloth. Therefore, privacy protection of the images and videos is necessary. Specifically, the detection and tracking methods should be used on the privacy protected images. For this purpose, there are some easy methods, such as blurring and pixelating, and they are often used in news programs etc. Because such methods just average pixel values, no important feature for the detection and tracking is left. Hence, the preprocessed images are unuseful. In order to solve this problem, we have proposed shuffle filter and a multi-view face tracking method with a genetic algorithm (GA). The filter protects the privacy by changing pixel locations, and the color information can be preserved. Since the color information is left, the tracking can be achieved by a basic template matching with histogram. Moreover, by using GA instead of sliding window when the subject in the image is searched, it can search more efficiently. However, the tracking accuracy is still low and the preprocessing time is large. Therefore, improving them is the purpose in this research. In the experiment, the improved method is compared with our previous work, CAMSHIFT, an online learning method, and a face detector. The results indicate that the accuracy of the proposed method is higher than the others.
{"title":"Evolutionary multi-view face tracking on pixel replaced image in video sequence","authors":"J. Sato, T. Akashi","doi":"10.1109/SOCPAR.2015.7492767","DOIUrl":"https://doi.org/10.1109/SOCPAR.2015.7492767","url":null,"abstract":"Nowadays, many computer vision techniques are applied to practical applications, such as surveillance and facial recognition systems. Some of such applications focus on information extraction from the human beings. However, people may feel psychological stress about recording their personal information, such as a face, behavior, and cloth. Therefore, privacy protection of the images and videos is necessary. Specifically, the detection and tracking methods should be used on the privacy protected images. For this purpose, there are some easy methods, such as blurring and pixelating, and they are often used in news programs etc. Because such methods just average pixel values, no important feature for the detection and tracking is left. Hence, the preprocessed images are unuseful. In order to solve this problem, we have proposed shuffle filter and a multi-view face tracking method with a genetic algorithm (GA). The filter protects the privacy by changing pixel locations, and the color information can be preserved. Since the color information is left, the tracking can be achieved by a basic template matching with histogram. Moreover, by using GA instead of sliding window when the subject in the image is searched, it can search more efficiently. However, the tracking accuracy is still low and the preprocessing time is large. Therefore, improving them is the purpose in this research. In the experiment, the improved method is compared with our previous work, CAMSHIFT, an online learning method, and a face detector. The results indicate that the accuracy of the proposed method is higher than the others.","PeriodicalId":409493,"journal":{"name":"2015 7th International Conference of Soft Computing and Pattern Recognition (SoCPaR)","volume":"124 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2015-11-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"115806772","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 : 2015-11-01DOI: 10.1109/SOCPAR.2015.7492777
Asmaa Hashem Sweidan, Nashwa El-Bendary, A. Hassanien, O. Hegazy, A. Mohamed
This paper presents a bio-inspired optimized classification approach for assessing water quality. As fish liver histopathology is a good biomarker for detecting water pollution, the proposed classification approach uses fish liver microscopic images in order to detect water pollution and determine water quality. The proposed approach includes three phases; preprocessing, feature extraction, and classification phases. Color histogram and Gabor wavelet transform have been utilized for feature extraction phase. The Machine Learning (ML) Support Vector Machines (SVMs) classification algorithm has been employed, along with the bio-inspired Gray Wolf Optimization (GWO) algorithm for optimizing SVMs parameters, in order to classify water pollution degree. Experimental results showed that the average accuracy achieved by the proposed GWO-SVMs classification approach exceeded 95% considering a variety of water pollutants.
{"title":"Water quality classification approach based on bio-inspired Gray Wolf Optimization","authors":"Asmaa Hashem Sweidan, Nashwa El-Bendary, A. Hassanien, O. Hegazy, A. Mohamed","doi":"10.1109/SOCPAR.2015.7492777","DOIUrl":"https://doi.org/10.1109/SOCPAR.2015.7492777","url":null,"abstract":"This paper presents a bio-inspired optimized classification approach for assessing water quality. As fish liver histopathology is a good biomarker for detecting water pollution, the proposed classification approach uses fish liver microscopic images in order to detect water pollution and determine water quality. The proposed approach includes three phases; preprocessing, feature extraction, and classification phases. Color histogram and Gabor wavelet transform have been utilized for feature extraction phase. The Machine Learning (ML) Support Vector Machines (SVMs) classification algorithm has been employed, along with the bio-inspired Gray Wolf Optimization (GWO) algorithm for optimizing SVMs parameters, in order to classify water pollution degree. Experimental results showed that the average accuracy achieved by the proposed GWO-SVMs classification approach exceeded 95% considering a variety of water pollutants.","PeriodicalId":409493,"journal":{"name":"2015 7th International Conference of Soft Computing and Pattern Recognition (SoCPaR)","volume":"31 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2015-11-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"125341164","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 : 2015-11-01DOI: 10.1109/SOCPAR.2015.7492786
Ran Li, Xuezhen Li, Takio Kurita
Local Binary Pattern (LBP) is known as one of the most effective local descriptors for image recognition. It is invariant to monotonic gray-scale changes of the image. Local neighborhood information is gathered for each pixel of the image, and a binary code is generated by comparing its value with the value of the center pixel. Then a histogram of binary code is created by counting up the occurrences of the different binary patterns. In this paper we propose an extension of the original LBP by using a soft thresholding function instead of the hard thresholding function using in the original LBP. Then we construct the histogram by voting the weights calculated depending on the distance between the extracted feature vector and the binary vectors. By using the proposed Soft LBP, we can extract information on the differences between the value of the center pixel and the value of the neighboring pixels. This means that the details of the textures can be included in the extracted features. To confirm the effectiveness of the proposed Soft LBP, we have performed the experiments on face recognition and face expression recognition. The results shows that the proposed Soft LBP gives better recognition rates than the original LBP and and the co-occurrence of adjacent local binary pattern and is comparable with the Soft Histogram LBP.
{"title":"Soft local binary patterns","authors":"Ran Li, Xuezhen Li, Takio Kurita","doi":"10.1109/SOCPAR.2015.7492786","DOIUrl":"https://doi.org/10.1109/SOCPAR.2015.7492786","url":null,"abstract":"Local Binary Pattern (LBP) is known as one of the most effective local descriptors for image recognition. It is invariant to monotonic gray-scale changes of the image. Local neighborhood information is gathered for each pixel of the image, and a binary code is generated by comparing its value with the value of the center pixel. Then a histogram of binary code is created by counting up the occurrences of the different binary patterns. In this paper we propose an extension of the original LBP by using a soft thresholding function instead of the hard thresholding function using in the original LBP. Then we construct the histogram by voting the weights calculated depending on the distance between the extracted feature vector and the binary vectors. By using the proposed Soft LBP, we can extract information on the differences between the value of the center pixel and the value of the neighboring pixels. This means that the details of the textures can be included in the extracted features. To confirm the effectiveness of the proposed Soft LBP, we have performed the experiments on face recognition and face expression recognition. The results shows that the proposed Soft LBP gives better recognition rates than the original LBP and and the co-occurrence of adjacent local binary pattern and is comparable with the Soft Histogram LBP.","PeriodicalId":409493,"journal":{"name":"2015 7th International Conference of Soft Computing and Pattern Recognition (SoCPaR)","volume":"46 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2015-11-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"116943265","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：