I recently read your paper "Object Detection for Graphical User Interface: Old Fashioned or Deep Learning or a Combination?" and used your code which is from your GitHub repository. Using the pretrained weights and the train/val/test json files that you put on DropBox, I made the train/val/test txt files and tested the model which is named "PyTorch-YOLOv3" in your GitHub repository. I used the default parameters during the test and I was confused about the results: a very low precision(the mean is 0.03).Am I processing my data wrong?Could I get some help from you? I have sent an email to your Email address,but you don't reply to me, so I create this issue.
Hi Chen，
I recently read your paper "Object Detection for Graphical User Interface: Old Fashioned or Deep Learning or a Combination?" and used your code which is from your GitHub repository. Using the pretrained weights and the train/val/test json files that you put on DropBox, I made the train/val/test txt files and tested the model which is named "PyTorch-YOLOv3" in your GitHub repository. I used the default parameters during the test and I was confused about the results: a very low precision(the mean is 0.03).Am I processing my data wrong?Could I get some help from you? I have sent an email to your Email address,but you don't reply to me, so I create this issue.