Recent Advances of Generic Object Detection with Deep Learning: A Review
Object detection is an important and challenging problem in computer vision. It has been widely applied in many vision tasks, such as object tracking, image segmentation, action recognition, etc. With the rapid development of deep learning, more state-of-the-art object detection methods based on deep learning with some modifications have effectively improved the detection performance. This paper comprehensively reviews object detection methods in the recent five years based on deep learning from object detection framework, including significant advances of the backbone network, multi-scale learning, data augmentation. Finally, we investigate the performance of typical object detection algorithms on popular datasets MS-COCO, PASCAL-VOC, and point out the existing problem for further research.
PDF
