A Comprehensive Review of Moving Object Identification Using Background Subtraction in Dynamic Scenes

Authors

  • Pavan Kumar Tadiparthi, Sreenu Ponnada, K.S Jhansi , Pradeep Kumar Bheemavarapu, Amrutha Gottimukkala

Abstract

Generally, images are divided into the foreground part and the background part. The foreground detection method can distinguish foreground objects, e.g., Static or dynamic from the background. For background detection, there are many modeling methods that can effectively recognize the background part. These methods have many advantages and are used in many applications such as video analysis, in particular video surveillance, shopping centers, road traffic. With an excellent model that is used to recognize background parts, we can get the result to the basic truth precisely enough. In order to recognize the foreground part in an image, we have different modeling methods. Different challenges for these methods are dynamic background and lighting changes. The GMM, AGMM, and VIBE methods were analyzed and their performance in terms of quality and time metrics calculated.

Keywords- background subtraction; adaptive gaussian mixture model (AGMM); gaussian mixture model (GMM); VIBE.

Published

2021-03-17

Issue

Section

Articles