Face Recognition: Novel Comparison of Various Feature Extraction Techniques
Face Recognition: Novel Comparison of Various Feature Extraction Techniques
复制标题
人脸识别：各种特征提取技术的新颖比较
DOI：
10.1007/978-981-13-0761-4_110
复制
发表时间：
2018
期刊：
影响因子：
--
通讯作者：
Raman Sharma
中科院分区：
文献类型：
--
作者：
Yashoda Makhija;Raman Sharma
人脸识别系统当然可以识别照片中的一张脸。这包括提取图像的特征，然后识别它，尽管光照、表情、姿势、老化和变换(平移、旋转和缩放图像)是一项艰巨的任务。在接下来的研究论文中，对各种特征提取技术进行了全面的文献综述。为了进行全面的综述，我们对居住特征提取技术进行了分类，并对每种分类中的具体方法进行了详细描述。这些策略分为四个值得注意的类别，具体为基于特征的方法、基于外观的方法、基于模板的方法和基于零件的方法。我们工作的动机是无法对先前可用调查中的每个可行算法执行进行全面和直接的独立比较。经过对这些策略的深入研究，我们分析了各种特征提取技术为图像处理的各种应用提供了领先的结果。
Face recognition system certainly recognizes a face in a picture. This involves extracting features of an image and then recognizing it, despite lighting, expression, pose, aging, and transformations (translate, rotate, and scale image) which is a tough task. In the following research paper, a comprehensive literature review of various kinds of technologies for feature extraction is listed. To present a comprehensive review, we classify residing feature extraction technologies along with detailed description of specific approaches within each classification. These strategies are grouped into four noteworthy classifications, specifically, feature-based, appearance-based, template-based, and part-based approaches. The motivation for our work is the unavailability of comprehensive and direct independent comparison of each one of the feasible algorithm executions in the previously available survey. After considerable exploration of these strategies, we analyze that various feature extraction technologies provide leading results for various applications of image processing.