Updated the preprocessing script as well as readmes

README.md

@@ -18,6 +18,7 @@ trained and evaluated on LIVECell.
 ## How to access LIVECell
 
 All images in LIVECell are available following [this link](http://livecell-dataset.s3.eu-central-1.amazonaws.com/LIVECell_dataset_2021/images.zip)  (requires 1.3 GB). Annotations for the different experiments are linked below.
+To see a more details regarding benchmarks and how to use our models, see [this link](model/README.md).
 
 ### LIVECell-wide train and evaluate
 

code/preprocessing.py

@@ -34,7 +34,7 @@ def preprocess(input_image, magnification_downsample_factor=1.0):
     x = int(dims[1]*magnification_downsample_factor)
 
     #apply resizing image to account for different magnifications
-    output_image = cv2.resize(output_image, (x, y), interpolation = cv2.INTER_AREA)
+    output_image = cv2.resize(output_image, (x,y), interpolation = cv2.INTER_AREA)
 
     #clip here to regular 0-255 range to avoid any odd median filter results
     output_image[output_image > 255] = 255
@@ -50,3 +50,11 @@ def preprocess(input_image, magnification_downsample_factor=1.0):
     output_image = output_image.astype('uint8')
 
     return output_image
+
+def preprocess_fluorescence(input_image, bInvert=True, magnification_downsample_factor=1.0): 
+    #invert to bring background up to 128
+    img = (255-input_image)/2
+    if not bInvert:
+        img = 255-img
+    output_image = preprocess(img, magnification_downsample_factor=magnification_downsample_factor) 
+    return output_image

model/README.md

@@ -250,3 +250,7 @@ python train_net.py  --config-file livecell_config.yaml --eval-only MODEL.WEIGHT
 For further details on training, testing and inference, 
 visit the [centermask2](https://github.com/youngwanLEE/centermask2#evaluation) or 
 [detectron2-ResNeSt](https://github.com/chongruo/detectron2-ResNeSt/blob/resnest/GETTING_STARTED.md) docs
+
+## One-shot usage
+For LIVECell experiments with zero-shot learning of EVICAN and Cellpose the input images was preprocessed using the 
+preprocessing-script preprocessing.py found under the [code folder](code/preprocessing.py).