A Review on Deep Learning Techniques Applied to Object Detection

Object detection has received great attention in recent years. Traditional machine learning object detection pays more attention to extract interpretable features. Deep learning uses CNN for self-learning features, which greatly improves the accuracy. This article summarized and analyzed the existing deep learning object detection method from one-stage and two-stage object detection methods. The roadmap of deep learning target detection is comprehensively analyzed. We have compared the detection performance of typical deep learning target detection algorithms. Finally, aiming at the problems and development trend of machine vision, we make prediction and prospect for the application and development of object detection in the future.