Efficient algorithm for object detection using saliency feature in machine vision systems
Athira Mohan, Dhana Madhu, J Tayalekshmy, R P Aneesh, Juby Raju · 2019
Object detection and recognition is the important application of the machine vision systems. Saliency is modern modes which find the visible items from rushed scenes in daily routines. But mania forecast model in tangled surrounding is harsh to put into action. Contour based segmentation technique is now used in extracting these structures from scenes. But some objects cannot be easily identified with this method. In this paper, a novel technique is presented for the segmentation of the salient objects from both complex and simple scenes. The technique search the local and global information of object features in a random manner. Multilevel thresholding with YCbCr color space is used to segment the salient objects from the image. The depth of the salient object is also estimated to categorize the objects within the scenes. Texture analysis is also done to This scheme is successfully tested with MSRA, ECSSD dataset and acquired an accuracy of 94%.This algorithm is also tested in the detection of traffic sign board and in the tracking of human.