Analysis and Performance in Action Recognition

Abstract

Abstract

Human activity and movement examination has been explored under a few extensive research ventures around the world. Issues like foundation mess, fractional impediment, changes in scale, perspective, lighting and appearance drives the undertaking of perceiving human exercises from video arrangements or still pictures into a testing assignment. Numerous applications, including video reconnaissance frameworks, human-PC communication, and apply autonomy for human conduct portrayal, require a different action acknowledgment framework. In numerous PC vision applications, for example, video reconnaissance, human PC interface, video ordering and perusing, acknowledgment of signals, investigation of games, occasions and move choreography perceiving human activity is a key segment. In the real world environment, monocular videos are insufficient for the practical applicability of action recognition algorithms due to two problems, (1) 3D information is lost in monocular videos, (2) A single camera view usually cannot fully capture human action due to the occlusion. The study of human body motion perception by the human visual system was made possible by the use of the so called moving light displays (MLDS) in psychophysics. Early approaches for human action recognition focused on the low level motion analysis such as tracking and body posture analysis. The human action recognition task is challenging due to changes in the appearance of persons articulation is poses, changing backgrounds and camera movements.

Keywords: HMM, 2D, 3D, PC, MLDS

Cite this Article

Yashasvi Shah, Kothari AM. Analysis and Performance in Action Recognition. Journal of Image Processing & Pattern Recognition Progress. 2017; 4(1):      16–20p.