Self-adaptive deep learning detection model post-processing method

The invention discloses a self-adaptive deep learning detection model post-processing method. According to the method provided by the invention, the application does not need to pay attention to the structure and characteristics of the model, the model processing result can be quickly and conveniently obtained, and the deep learning visual detection and other models can be quickly applied to industrial defect detection, automatic driving, security and other fields. After the input is adjusted and the model structure is iterated, the model can be quickly online. Post-processing logic of the model does not need to be concerned and adjusted, application development debugging workload is reduced, and application development efficiency is greatly improved. The threshold of applying the deep learning model is reduced, and application and hardware development engineers can quickly apply artificial intelligence and the deep learning model to solve application problems without deeply knowing the model structure. 本发明公开