A Nobel Approach to Detect Mysterious Action at Real-Time from Object Speed and Background Subtraction by using Incremental Tensor Subspace Learning

Contenu principal de l'article

Kusum Yadav , Afraa Sayah Alshammari , Aljawharah Almuhana, Anurag Jain


Abnormal action detection is a very important process for avoiding crime happening in various areas where there is no human resource. From surveillance videos abnormality like chasing, murders, races, fights, accidents, etc. should be detected automatically and intimation will be given to respective personalities. The goal of the proposed work is to recognize mysterious action detection by speed and displacement analysis in image processing. In the proposed work, authors have used a combination of motion detection along with speed and displacement measurement using various image analysis techniques implemented in MATLAB simulation environment. Motion detection is determined by Incremental and Multi-feature Tensor Subspace Learning Applied for Background Modelling and Subtraction. After Background subtraction, the procedure of moving object detection comes into the action. The threshold has been fixed for normal speed and displacement exceeding this shows mysterious action happening and the corresponding video has been popped up along by saving the frames with the action noted.


Renseignements sur l'article