Review of person re‐identification techniques

Abstract

Person re-identification across different surveillance cameras with disjoint fields of view has become one of the most interesting and challenging subjects in the area of intelligent video surveillance. Although several methods have been developed and proposed, certain limitations and unresolved issues remain. In all of the existing re-identification approaches, feature vectors are extracted from segmented still images or video frames. Different similarity or dissimilarity measures have been applied to these vectors. Some methods have used simple constant metrics, whereas others have utilised models to obtain optimised metrics. Some have created models based on local colour or texture information, and others have built models based on the gait of people. In general, the main objective of all these approaches is to achieve a higher-accuracy rate and lowercomputational costs. This study summarises several developments in recent literature and discusses the various available methods used in person re-identification. Specifically, their advantages and disadvantages are mentioned and compared.

Review of Current Methods for Re-Identification in Computer Vision.

“…Camera calibration [5] or different types of cameras can increase the difficulty in ReID a person. A change in a camera by either; height, angle, brand, lens, and even placement can have adverse effects on subject identification [6,4]. This issue is known as an intra-class variation.…”

“…The gait analysis is one of the least favorite methods for ReID because of the high possibility of the change in rotation, scale, and angle of view of the target. These alterations can make matching the gate of a person nearly impossible [4,19]. Feature extraction based on the target's physical appearance can be some of the more straightforward methods to implement and give decent results depending on the environmental circumstances [20].…”

“…The process of ReID focuses on correctly matching a target from either a sample or a dataset and find the target in another image regardless of any change in position or location. These changes can range from altered lighting, the target's position or pose, the scale of the target, background changes, and camera rotations or orientation [4].…”

A Survey on Face Detection and Person Re-identification

“…The person re-identification task is challenging because the images of person captured by one camera May vary significantly from other or the images of the same person May be viewed differently by the same camera at a later time due to illumination condition, pose changes and scale. In [23] the issues of person re identification is divided into inter-camera and intra-camera. inter-camera issues include illumination changes, disjoint fields of view, different entry poses in different cameras, similarity of clothing; rapid changes in person's clothing; and blurred images.…”

“…Intra-camera issues include background illumination changes; low resolution of CCTVs and occlusion in frames. Ali Saghafi [23] grouped the person re-identification methods into two groups. The first one is signature based methods which extract color, shape, texture etc.…”