Author : Chala Simon and Shilpa Gite

A scene is a real-world view of the environment that contains different surfaces and objects, organised in a meaningful way. Since low-level features obtained from the video stream are insufficient and limited to describe the scene, it’s difficult for the classifier to identify the objects on the scene due to less information about the image. Image fusion is the procedure of combining multiple image information into one to produce more steady and useful information. Using high-level semantic descriptor can also help the classifier to perform classification easily. The aspects explored in this paper to use fused information of the low-level features along with a high-level semantic descriptor of an image sequence from vehicle dashboard-mounted camera to a better understanding of the scene. Here we have proposed to develop vision-based road awareness and object tracking for driving assistance purpose.

Keywords: ADAS, Context-enhanced, Computer vision, feature extraction, multi-resolution analysis based image fusion

Article published in International Journal of Current Engineering and Technology, Vol.7, No.6 (Nov/Dec 2017)