Added app.py and requirements.txt

app.py

@@ -0,0 +1,97 @@
+import gradio as gr
+import pydicom
+import numpy as np
+import cv2
+import pandas as pd
+
+def load_dicom(file):
+    """Load a DICOM file and extract pixel data with rescale slope and intercept applied."""
+    dicom_data = pydicom.dcmread(file)
+    image = dicom_data.pixel_array.astype(np.float32)
+    rescale_slope = getattr(dicom_data, "RescaleSlope", 1)
+    rescale_intercept = getattr(dicom_data, "RescaleIntercept", 0)
+    image = (image * rescale_slope) + rescale_intercept
+    pixel_spacing = getattr(dicom_data, "PixelSpacing", [1, 1])
+    return image, pixel_spacing
+
+def analyze_dicom(file, circle_diameter):
+    """
+    Analyze the DICOM file:
+    - Load image and pixel spacing
+    - Display the image
+    - Analyze a circular region of interest
+    """
+    try:
+        # Load DICOM file
+        image, pixel_spacing = load_dicom(file)
+        pixel_size_mm = float(pixel_spacing[0])  # Assuming square pixels
+        
+        # Normalize image for display
+        image_display = cv2.normalize(image, None, 0, 255, cv2.NORM_MINMAX).astype(np.uint8)
+        
+        # Convert to RGB for overlay purposes
+        image_rgb = cv2.cvtColor(image_display, cv2.COLOR_GRAY2RGB)
+
+        # Analyze center of the image as a circular ROI
+        h, w = image.shape
+        center_x, center_y = w // 2, h // 2
+
+        # Create a circular mask
+        mask = np.zeros_like(image, dtype=np.uint8)
+        y_indices, x_indices = np.ogrid[:image.shape[0], :image.shape[1]]
+        distance_from_center = np.sqrt((x_indices - center_x) ** 2 + (y_indices - center_y) ** 2)
+        mask[distance_from_center <= circle_diameter / 2] = 1
+
+        # Extract pixel values within the circle
+        pixels = image[mask == 1]
+
+        # Calculate metrics
+        area_pixels = np.sum(mask)
+        area_mm2 = area_pixels * (pixel_size_mm**2)
+        mean = np.mean(pixels)
+        stddev = np.std(pixels)
+        min_val = np.min(pixels)
+        max_val = np.max(pixels)
+
+        # Save the result
+        results = {
+            "Area (mm²)": f"{area_mm2:.3f}",
+            "Mean Intensity": f"{mean:.3f}",
+            "Standard Deviation": f"{stddev:.3f}",
+            "Minimum Intensity": f"{min_val:.3f}",
+            "Maximum Intensity": f"{max_val:.3f}",
+        }
+
+        # Draw the circle on the image
+        cv2.circle(image_rgb, (center_x, center_y), int(circle_diameter / 2), (0, 255, 0), 2)
+        
+        # Convert the image back to BGR for display
+        annotated_image = cv2.cvtColor(image_rgb, cv2.COLOR_RGB2BGR)
+
+        return annotated_image, pd.DataFrame([results]).to_csv(index=False)
+    except Exception as e:
+        return None, f"Error: {str(e)}"
+
+# Define the Gradio interface
+inputs = [
+    gr.File(label="Upload DICOM File (.IMA or .dcm)", type="file"),
+    gr.Number(label="Circle Diameter (in pixels)", value=50, precision=1),
+]
+
+outputs = [
+    gr.Image(label="Annotated DICOM Image"),
+    gr.File(label="Analysis Results (CSV)"),
+]
+
+# Launch the app
+gr.Interface(
+    fn=analyze_dicom,
+    inputs=inputs,
+    outputs=outputs,
+    title="DICOM Analyzer Tool",
+    description=(
+        "Upload a DICOM file and analyze a circular region of interest (ROI). "
+        "The tool calculates metrics such as mean intensity, standard deviation, and area in mm²."
+    ),
+    theme="compact",
+).launch()

requirements.txt

@@ -0,0 +1,5 @@
+gradio
+numpy
+pandas
+opencv-python
+pydicom