search

nv-nguyen / cnos

[ICCV 2023 R6D] PyTorch implementation of CNOS: A Strong Baseline for CAD-based Novel Object Segmentation based on Segmenting Anything and DINOv2
MIT License
189 stars 21 forks source link

Help Needed: Improving Prediction Accuracy on Custom Dataset #23

Open EwertzJN opened 1 month ago

EwertzJN commented 1 month ago

Thank you for your great work — so far I have found it very helpful! To run your work on a dataset I created, I wrote the attached script. For the most part, I'm getting good results, but I'm also having problems with a few predictions (missing detections, incorrect class assignments and incorrect detections), as the following visualizations show, for example:

masks masks masks masks masks

I would be very grateful if you could, based on your experience in using the model, give me some hints on what changes I could perhaps try to improve the results a bit more. Thank you very much!

EwertzJN commented 1 month ago

It didn't solve my problem, but regardless of that, could it be that it was forgotten here to apply the RGB transformation to the templates? I noticed this in my script, which is partly inspired by inference_custom.py. I have opened a pull request about this. If I am wrong, please feel free to close it.