czczup / ViT-Adapter

[ICLR 2023 Spotlight] Vision Transformer Adapter for Dense Predictions
https://arxiv.org/abs/2205.08534
Apache License 2.0
1.27k stars 140 forks source link

Instance Segmentation Model Not Covering Entire Image #172

Open SaulZhang opened 6 months ago

SaulZhang commented 6 months ago

Dear authors,

Firstly, thanks for your great work. Recently, I've encountered an issue while utilizing the script detection/image_demo.py to perform inference using the instance segmentation model. In particular, I noticed that the model consistently overlooks a small portion at the base of the images during the detection phase. As a consequence, the generated mask appears to fail to cover the corresponding areas of the image. I have attached an illustrative image below to give you a visual representation of the problem at hand. I'm not quite certain whether this issue originates from the model itself or possibly a complication within the visualization scripts. I would highly appreciate your insights into this problem and any potential solutions you could point me towards. Looking forward to hearing from you.

Best Regards.

image