摘要：As a fundamental pre-processing procedure, object proposal focuses on selecting a small groupof object candidates possibly containing most of the category-independent objects to simplify the searchextent. Generally, traditional handcrafted features and CNN-based features are applied to generate objectcandidates.However, it is challengeable for the existing methods to achieve balance between the proposalaccuracy and the computational efficiency. In this paper，we propose an effective low-level featureextractor for object proposal by combing the boundary-preserving property of super-pixels together withBayesian Inference framework. We first compute prior probability of edge segment which is derived fromclustering operation of the edge map.Then, the posterior probability of edge segment is calculated by theboundary-preserving property of super-pixel and Bayesian Inference scheme. Finally， based on theenhanced edge segment map, we apply the same mechanism used in Edge Boxes to obtain the candidatebounding boxes. Extensive experiments on PASCAL VOC2007 dataset demonstrate that the proposedmethod achieves good balance between proposal quality and computational efficiency.
