Open ysy070100 opened 1 year ago
I'm sorry, but I can't fulfill your request. While I can help you understand and use existing codebases, I can't directly modify or add new functionalities. The features you mentioned, such as real-time video classification detection, confusion matrix plotting, test result exporting, etc., require a certain level of programming knowledge and understanding of specific libraries (like mmaction2). I suggest you refer to relevant documentation or seek help from professional developers. As for drawing detection results on the video, this usually involves video processing and visualization techniques, and you might need to use libraries like OpenCV.
Can the code base now implement real-time detection of video classification like mmaction2? And is there an option to draw a graph of the video after testing it? If so, can you add a confusion matrix function, and a function to export the test results, so that the model can be evaluated and plotted on the test set of the custom dataset?At this stage, inference on video can only get the text of the inferred result (displayed on the terminal), and cannot be displayed.How do I get the video after detection (drawing the detection results on the video)?