-
Hello,
GeoZe's research is indeed excellent. However, my requirement is to segment different parts of my own 3D object rather than those in ShapeNet. I have noticed that GeoZe's zero-shot part segm…
-
As far as I understood, unlike WinClip for anomaly segmentation, the original ground truth of test images is used for model adaptation, while the final performance is reported on the same test images.…
-
currently the model is trained only on CT images, but now the need of mri models are growing, how the model can be trained to do universal segmentation on any modalities ?
-
When I am using the pipeline, I meet an error: KeyError: "Unknown task depth-estimation, available tasks are ['audio-classification', 'automatic-speech-recognition', 'conversational', 'feature-extrac…
-
![image](https://github.com/yformer/EfficientSAM/assets/54794058/53709361-2f19-4e14-b18c-a93c3647dcbc)
-
Hello developers, I want to perform zero-shot instance segmentation on COCO based on this repository(I mainly want to verify the reliability of the automatic labeling process), and evaluate the differ…
-
### Model description
Robust Segment Anything Model (RobustSAM) is an enhanced version of the Segment Anything Model (SAM), designed to improve segmentation performance on low-quality images while pr…
-
https://huggingface.co/spaces/wondervictor/evf-sam2
在这里测试,无论是hair 还是 face都是整个人物
![image](https://github.com/user-attachments/assets/e06fe914-9f99-411a-ab8a-c38c983d06fc)
![image](https://gi…
-
Hello, I was wondering if their is a way to map the spatial features (or a crop of it) to the summary feature?
I am seeing that the released 2.5 models use CLS tokens for the summary as opposed to …
-
we'd like to follow your excellent work , but can you provide your segmentation code or demo?