Recognition of salient object

Abstract

Salient region detection is essential theme in computer vision. This paper demonstrates the application of salient region detection for object recognition and classification. The different visual cues are used for salient object detection. The local contrast and compactness visual cues are complementally to each other. The salient regions are not correctly suppressed by compactness cues but local contrast can effectively recover. The bottom-up technique is used for recognition of salient object. The saliency map is getting from compactness and local contrast map. After detection of salient region or objects we will extend our approach towards recognition or of some special objects or shapes. For that we will work on some geometrical features.