Update app.py

app.py

@@ -27,7 +27,7 @@ GLEASON_LABELS = {
 
 }
 BACH_LABELS = {0: "Benign",
-               1: "InSitu",
+               1: "In Situ",
                2:"Invasive",
                3: "Normal"}
 CRC_LABELS = {
@@ -194,13 +194,13 @@ def predict_class(image, linear, dataset):
 breakhis = gr.Interface(
     fn=predict_breakhis,
     inputs=gr.Image(type="filepath", label="Upload Breast Histopathology Image"),
-    outputs=gr.Label(num_top_classes=4, label="Tumor Type Prediction"),
-    title="BreakHis Breast Tumor Classification",
+    outputs=gr.Label(num_top_classes=4, label="BreakHis Breast Cancer Classification"),
+    title="BreakHis Breast Cancer Classification",
     description="""
-    Upload a breast histopathology image to predict the tumor type. Your image must be at 40X magnification, and ideally between 224x224 and 700x460 resolution. Do not otherwise modify your image.
+    Upload a breast histopathology image to predict the breast cancer subtype. Your image must be at 40X magnification, and ideally between 224x224 and 700x460 resolution. Do not otherwise modify your image.
 
-    This model uses a custom-trained DinoV2 foundation model for pathology images
-    with a linear classifier for BreakHis tumor classification.
+    This demo uses a custom-trained DINOv2 foundation model for pathology images called [OpenMidnight](https://sophont.med/blog/openmidnight)
+    with a linear classifier for BreakHis breast cancer classification.
 
     **Tumor Types:**
     - **Benign tumors:** Tubular Adenoma (TA), Fibroadenoma (F)
@@ -208,6 +208,8 @@ breakhis = gr.Interface(
 
     These 4 classes were selected from the full BreakHis dataset as they have sufficient patient counts (≥7 patients) for robust evaluation.
     For this particular demo, images *must* be one of the sample classes - unsupported classes will yield confusing and/or useless results.
+
+    This demonstration is for illustrative purposes only and should not be used for diagnostic/clinical purposes.
     """,
     examples=["./SOB_B_TA-14-13200-40-001.png",
         "./SOB_M_MC-14-10147-40-001.png",
@@ -219,17 +221,19 @@ breakhis = gr.Interface(
 gleason = gr.Interface(
     fn=predict_gleason,
     inputs=gr.Image(type="filepath", label="Upload Prostate Cancer Image"),
-    outputs=gr.Label(num_top_classes=4, label="Gleason Tumor Type Prediction"),
-    title="Gleason Prostate Tumor Classification",
+    outputs=gr.Label(num_top_classes=4, label="Gleason Grading"),
+    title="Gleason Grading",
     description="""
     Upload a prostate cancer image to predict the tumor type. Your image must be at 40X magnification, and ideally between 224x224 and 750x750 resolution. Do not otherwise modify your image.
 
-    This model uses a custom-trained DinoV2 foundation model for pathology images
-    with a linear classifier for gleason tumor classification.
+    This demo uses a custom-trained DINOv2 foundation model for pathology images called [OpenMidnight](https://sophont.med/blog/openmidnight)
+    with a linear classifier for Gleason grading.
 
     Images are classified as benign, Gleason pattern 3, 4 or 5.
 
     For this particular demo, images *must* be one of the sample classes - unsupported classes will yield confusing and/or useless results.
+
+    This demonstration is for illustrative purposes only and should not be used for diagnostic/clinical purposes.
     """,
     examples=["./ZT111_4_A_1_12_patch_13_class_2.jpg",
         "./ZT204_6_A_1_10_patch_10_class_3.jpg",
@@ -242,16 +246,18 @@ crc = gr.Interface(
     fn=predict_crc,
     inputs=gr.Image(type="filepath", label="Upload Colorectal Cancer Image"),
     outputs=gr.Label(num_top_classes=9, label="CRC Tumor Type Prediction"),
-    title="Colorectal Tumor Classification",
+    title="Colorectal Cancer Tissue Classification",
     description="""
-    Upload a colorectal cancer image to predict the tumor type. Your image must be at 20X magnification, and ideally at 224x224. Do not otherwise modify your image.
+    Upload a colorectal cancer image to predict the tissue class. Your image must be at 20X magnification, and ideally at 224x224. Do not otherwise modify your image.
 
-    This model uses a custom-trained DinoV2 foundation model for pathology images
-    with a linear classifier for colorectal tumor classification.
+    This demo uses a custom-trained DINOv2 foundation model for pathology images called [OpenMidnight](https://sophont.med/blog/openmidnight)
+    with a linear classifier for colorectal cancer tissue classification.
 
     The tissue classes are: Adipose (ADI), background (BACK), debris (DEB), lymphocytes (LYM), mucus (MUC), smooth muscle (MUS), normal colon mucosa (NORM), cancer-associated stroma (STR) and colorectal adenocarcinoma epithelium (TUM)
 
     For this particular demo, images *must* be one of the sample classes - unsupported classes will yield confusing and/or useless results.
+
+    This demonstration is for illustrative purposes only and should not be used for diagnostic/clinical purposes.
     """,
     examples=["./ADI-TCGA-AAICEQFN.png",
         "./BACK-TCGA-AARRNSTS.png",
@@ -263,17 +269,19 @@ crc = gr.Interface(
 bach = gr.Interface(
     fn=predict_bach,
     inputs=gr.Image(type="filepath", label="Upload Cancer Image"),
-    outputs=gr.Label(num_top_classes=4, label="Bach Tumor Type Prediction"),
-    title="Tumor Classification",
+    outputs=gr.Label(num_top_classes=4, label="BACH Breast Cancer Classification"),
+    title="BACH Breast Cancer Classification",
     description="""
     Upload a breast cancer image to predict the tumor type. Your image must be at 20X magnification, and ideally between 224x224 and 1536x2048 resolution. Do not otherwise modify your image.
 
-    This model uses a custom-trained DinoV2 foundation model for pathology images
+    This demo uses a custom-trained DINOv2 foundation model for pathology images called [OpenMidnight](https://sophont.med/blog/openmidnight)
     with a linear classifier for tumor classification.
 
-    Images are classified as benign, normal, invasive, inSitu
+    Images are classified as benign, normal, invasive, in-situ.
 
     For this particular demo, images *must* be one of the sample classes - unsupported classes will yield confusing and/or useless results.
+
+    This demonstration is for illustrative purposes only and should not be used for diagnostic/clinical purposes.
     """,
     examples=["./b001.png",
         "./n001.png",
@@ -285,17 +293,20 @@ bach = gr.Interface(
 bracs = gr.Interface(
     fn=predict_bracs,
     inputs=gr.Image(type="filepath", label="Upload Cancer Image"),
-    outputs=gr.Label(num_top_classes=7, label="Bracs Tumor Type Prediction"),
-    title="Tumor Classification",
+    outputs=gr.Label(num_top_classes=7, label="BRACS Tumor Subtyping"),
+    title="BRACS Tumor Subtyping",
     description="""
     Upload a breast cancer image to predict the tumor type. Your image must be at 40X magnification. Do not otherwise modify your image.
 
-    This model uses a custom-trained DinoV2 foundation model for pathology images with a linear classifier for tumor classification.
+    This demo uses a custom-trained DINOv2 foundation model for pathology images called [OpenMidnight](https://sophont.med/blog/openmidnight)
+    with a linear classifier for tumor classification.
 
     Images are classified as Normal, Pathological Benign, Usual Ductal Hyperplasia, Flat Epithelial Atypia, 
         Atypical Ductal Hyperplasia, Ductal Carcinoma In Situ, Invasive Carcinoma
 
     For this particular demo, images *must* be one of the sample classes - unsupported classes will yield confusing and/or useless results.
+
+    This demonstration is for illustrative purposes only and should not be used for diagnostic/clinical purposes.
     """,
     examples=[
         ],  # You can add example image paths here