Closed k-chrispens closed 2 months ago
It would be useful to think of better methods for automatically selecting point prompts - both click methods I implemented are quite basic. I've seen decent results for my implementation in ultrasound segmentation though!
what a great job!I also think the method based on image thresholding is a really good way to click points.Its more like human to click points.
Addresses #22 #62 and somewhat #64
Makes a new dataset class (
Dataset_Union_ALL_Infer
) that doesn't return labels for using during inference.Implements new click methods (
get_next_click3D_torch_no_gt
andget_next_click3D_torch_no_gt_naive
) that allow naive inference - these generate point prompts based on image thresholding and random selection of points in the volume.The
infer_sequence.py
andinfer_sequence.sh
are built to use these to allow inference with no label mask.