Please use this identifier to cite or link to this item:
http://hdl.handle.net/10603/341983
Title: | Multi level region growing approach for efficient face tracking in video surveillance with invariant features using tri view binary patterns and |
Researcher: | Vivek Yoganand, A |
Guide(s): | Celine Kavida, A And Rukmanidevi, D |
Keywords: | Video surveillance Security Image processing |
University: | Anna University |
Completed Date: | 2019 |
Abstract: | The modern information informatics world provides various solution to the human society. The same has been utilized for challenging the security of more sensitive information of many organizations. The security of many organization has been implied using many biological features, among them the face features has become more popular in this era. The face features has been used not only in the security enforcement but also being used in many other applications. The face detection has been performed in different situations and from different face features. The face detection from face image has been performed in many ways. The recent problem being challenged by the researcher is identifying multiple faces from the same frame of video. The surveillance video has been used for face detection and tracking to enforce security in the organizations and for many other applications. Number of features has been used for face detection like color, shape and fractal features. Different methods have been discussed earlier which uses the above mentioned features. Still they suffer to achieve the higher performance in face detection and tracking, because the face feature present in the video frame would be captured in different illumination, occlusion and multiple views. This introduces higher challenge to the existing algorithms in identifying the face component. Towards the improvement of face detection in surveillance video. A region growing approach with neural network has been presented. The region growing technique is adapted to improve the performance of segmentation to identify the face features. The method extracts the color and shape features from the input frame. The method extracts the color features and the features have been trained under the neural network. The ABCalgorithm has been used for classification and the method improves the performance of classification and detection accuracy. However the accuracy of the region growing technique could be improved by adapting multiview features because of the frames would |
Pagination: | xix,210 p. |
URI: | http://hdl.handle.net/10603/341983 |
Appears in Departments: | Faculty of Information and Communication Engineering |
Files in This Item:
Items in Shodhganga are licensed under Creative Commons Licence Attribution-NonCommercial-ShareAlike 4.0 International (CC BY-NC-SA 4.0).
Altmetric Badge: