@@ -52,21 +52,25 @@ The strength of our approach is we can match supervised model performance with p
5252
5353![ FIG1 (1)] ( https://github.com/user-attachments/assets/0d970b45-79ff-4c58-861f-e1e7dc9abc65 )
5454
55- ** Figure 1. Performance of 3D Semantic and Instance Segmentation Models.**
56- ** a:** Raw mesoSPIM whole-brain sample, volumes and corresponding ground truth labels from somatosensory (S1) and visual (V1) cortical regions.
55+ ** Figure 1. Performance of 3D Semantic and Instance Segmentation Models.**
56+ ** a:** Raw mesoSPIM whole-brain sample, volumes and corresponding ground truth labels from somatosensory (S1) and visual (V1) cortical regions.
5757** b:** Evaluation of instance segmentation performance for baseline
5858thresholding-only, supervised models: Cellpose, StartDist, SwinUNetR, SegResNet, and our self-supervised model WNet3D over three data subsets.
5959F1-score is computed from the Intersection over Union (IoU) with ground truth labels, then averaged. Error bars represent 50% Confidence Intervals
60- (CIs).
61- ** c:** View of 3D instance labels from supervised models, as noted, for visual cortex volume in b evaluation.
60+ (CIs).
61+ ** c:** View of 3D instance labels from supervised models, as noted, for visual cortex volume in b evaluation.
6262** d:** Illustration of our WNet3D architecture showcasing the dual 3D U-Net structure with our modifications.
6363
6464
6565## News
6666
67- ** New version: v0.2.1 **
67+ ** New version: v0.2.2 **
6868
69- - v0.2.1:
69+ - v0.2.2:
70+ - Updated Colab notebooks for training and inference!
71+ - New models available in inference demo notebook
72+ - CRF optional post-processing adjustments
73+ - v0.2.2:
7074 - Updated plugin default behaviors across the board to be more readily applicable to demo data
7175 - Threshold value in inference is now automatically set by default according to performance on demo data on a per-model basis
7276 - Added a grid search utility to find best thresholds for supervised models
0 commit comments