Recent Trends in Person Re-identification from Videos

The purpose of Person re-identification is to locate a targeted person from a given image or video obtained from different camera feeds. There is a lot of research going on in the field of computer vision especially on Person re-identification. The conventional method used were template matching techniques. Due to the availability of high-end computing systems and large datasets, the person re-identification system works very well as the features extracted after the images are pre-processed, which almost solve the problems related to illumination variation, pose variation and occlusion. This paper discusses the related work and recent techniques used for person re-identification such as Histogram of Gradients, Scale Invariant Feature Transform, Local Binary Pattern, and deep learning methods.