Recognition of Partially Occluded Objects: A Faster R-CNN Approach
This paper addresses the problem of recognizing possible objects for use in partially occluded object recognition. To enable the use of more expensive features and classifiers and thereby progress beyond the state-of-the-art, a region proposal network (RPN) which shares full-image convolutional feature with detector network is needed. We aim to improve the recognition of partially occluded objects in the cluttered scene. In this paper, we build our approach based on the recent state-of-the-art Faster R-CNN to increase the recognition capability in partially occluded object recognition. We evaluate our approach on the real-time object recognition in the robotics application. The results demonstrate the effectiveness of our proposed approach for robust object recognition in real-time applications.
Keywords - Partially occluded, object recognition, Faster R-CNN, cluttered scene.