Create app.py

app.py

@@ -0,0 +1,57 @@
+import gradio as gr
+import cv2
+import numpy as np
+from transformers import DetrForObjectDetection, DetrImageProcessor
+import librosa
+import matplotlib.pyplot as plt
+
+# Load pre-trained model for ball detection
+processor = DetrImageProcessor.from_pretrained("facebook/detr-resnet-50")
+model = DetrForObjectDetection.from_pretrained("facebook/detr-resnet-50")
+
+def process_drs_video(video_path, audio_path=None):
+    # Extract frames from video
+    cap = cv2.VideoCapture(video_path)
+    frames = []
+    while cap.isOpened():
+        ret, frame = cap.read()
+        if not ret:
+            break
+        frames.append(frame)
+    cap.release()
+
+    # Detect ball in frames (simplified)
+    ball_positions = []
+    for frame in frames:
+        inputs = processor(images=frame, return_tensors="pt")
+        outputs = model(**inputs)
+        # Process outputs to get ball coordinates (custom logic needed)
+        ball_positions.append([100, 200])  # Placeholder
+
+    # Predict trajectory (simplified linear model)
+    trajectory = np.array(ball_positions)  # Replace with actual model
+
+    # Audio processing for edge detection
+    if audio_path:
+        y, sr = librosa.load(audio_path)
+        onset_env = librosa.onset.onset_strength(y=y, sr=sr)
+        edge_detected = np.max(onset_env) > 0.5  # Simplified threshold
+
+    # LBW decision logic
+    decision = "Out" if trajectory[-1][1] < 300 and not edge_detected else "Not Out"
+
+    # Visualize trajectory
+    plt.plot(trajectory[:, 0], trajectory[:, 1], 'r-')
+    plt.title(f"DRS Decision: {decision}")
+    plt.savefig("trajectory.png")
+
+    return decision, "trajectory.png"
+
+# Gradio interface
+iface = gr.Interface(
+    fn=process_drs_video,
+    inputs=[gr.Video(label="Upload Delivery Video"), gr.Audio(label="Upload Stump Mic Audio")],
+    outputs=[gr.Textbox(label="Decision"), gr.Image(label="Trajectory")],
+    title="Cricket DRS Demo"
+)
+iface.launch()