Closed ChristianLus closed 3 years ago
Okay so it looks like everything up to the Matlab evaluation code is working for me. Should I place the Pedestron_Result/ folder next to the ResultsEval folder when I bring them into Matlab? I suppose don't quite understand where the evaluation code is trying to look for the detection results.
Thank you again for your quick response.
Did you follow the instructions in this file particularly for Caltech
point 4 and 5 ?
I followed them as well as I could. I will give it another try and let you know if it is able to run the files. I apologize for any confusion.
Just see how the Matlab code is expecting your paths to be, its most definitely a path issue.
Hello, I wanted to let you know that I was able to figure it out. I'm not sure what I was doing wrong at first but now the evaluation code is running just fine. Thank you for your help!
Hello, I am fairly new to deep learning as well as github in general so I apologize if my questions are not clear.
When I evaluate on Caltech, the output is a a single .json file with detections for 4,024 images. When I go to convert these with the convert to .txt, I was able to run the conversion code only if I called the convert_result(results_jsons) directly rather than using p.map().
That being said, it creates a text file for each of the video files (V000.txt-V010.txt, depending on the set). Is this correct, or should the output be different?
When I bring these .txt into matlab, how should I form the hierarchy of these results so that it is able to read everything properly?
I am experiencing issues where the matlab code cannot find the detections and seems to be wanting a detection file per frame. Am I doing something wrong or am I simply misinterpreting the hierarchy?
Thank you for your time and I am happy to provide any additional information that may help you or anyone else.