|
Over the years, with the continuous improvement and development of deep learning algorithm, especially in combination with object detection, a breakthrough has been made. Therefore, the third generation of intelligent monitoring technology combined with computer vision has also achieved great development, and has played a very important role in the real-time traffic,banking and other fields. The Single Shot MultiBox Detector (SSD) is a multi-target fast detection algorithm that directly predicts object classes and bounding boxes. Because of its excellent overall performance, it is one of the mainstream frameworks for object detection so far and is widely applied Image recognition.Through careful analysis of the object detection of deep learning, this paper uses SSD to detect the object that cheating in the field of the examination\'s surveillance video and improves the output of detection results by combining triplet loss.The experimental results show that detection results has been achieved quickly and accurately under different scenarios by this method. |
|
Keywords:Pattern recognition; Deep learning;Object detection;SSD |
|