Merge pull request #105 from computational-cell-analytics/multi-object-point-prompts

Minor changes

Author: constantinpape

.github/workflows/doc_env.yaml .github/doc_env.yaml.github/workflows/doc_env.yaml renamed to .github/doc_env.yaml

@@ -24,7 +24,7 @@ jobs:
           activate-environment: sam
           mamba-version: "*"
           auto-update-conda: true
-          environment-file: .github/workflows/doc_env.yaml
+          environment-file: .github/doc_env.yaml
           python-version: ${{ matrix.python-version }}
           auto-activate-base: false
         env:

.github/workflows/build_docs.yaml

@@ -45,11 +45,12 @@ If you are using this repository in your research please cite
 
 ## Related Projects
 
-There are two other napari plugins build around segment anything:
+There are a few other napari plugins build around Segment Anything:
 - https://github.com/MIC-DKFZ/napari-sam (2d and 3d support)
 - https://github.com/JoOkuma/napari-segment-anything (only 2d support)
+- https://github.com/hiroalchem/napari-SAM4IS
 
-Compared to these we support more applications (2d, 3d and tracking), and provide finetuned models for microscopy data.
+Compared to these we support more applications (2d, 3d and tracking), and provide finetuning methods and finetuned models for microscopy data.
 [WebKnossos](https://webknossos.org/) also offers integration of SegmentAnything for interactive segmentation.
 
 

README.md

@@ -2,18 +2,29 @@
 from micro_sam.sample_data import fetch_image_series_example_data
 
 
-def series_annotation():
-    """Annotate a series of images. Example runs for three different example images.
+def series_annotation(use_finetuned_model):
+    """Annotate a series of images. Example runs for three different images.
     """
+
+    if use_finetuned_model:
+        embedding_path = "./embeddings/series-embeddings-vit_h_lm"
+        model_type = "vit_h_lm"
+    else:
+        embedding_path = "./embeddings/series-embeddings"
+        model_type = "vit_h"
+
     example_data = fetch_image_series_example_data("./data")
     image_folder_annotator(
-        example_data, "./data/series-segmentation-result", embedding_path="./embeddings/series-embeddings",
-        pattern="*.tif", model_type="vit_b"
+        example_data, "./data/series-segmentation-result", embedding_path=embedding_path,
+        pattern="*.tif", model_type=model_type
     )
 
 
 def main():
-    series_annotation()
+    # whether to use the fine-tuned SAM model
+    # this feature is still experimental!
+    use_finetuned_model = False
+    series_annotation(use_finetuned_model)
 
 
 if __name__ == "__main__":