Create app.py

app.py

@@ -0,0 +1,73 @@
+import gradio as gr
+import cv2
+import numpy as np
+import torch
+from transformers import ViTForImageClassification, ViTImageProcessor
+import mediapipe as mp
+
+# Load pretrained model (ASL word classifier)
+model_name = "your_pretrained_asl_model"  # Replace with actual model
+model = ViTForImageClassification.from_pretrained(model_name)
+processor = ViTImageProcessor.from_pretrained(model_name)
+
+# MediaPipe Hands setup
+mp_hands = mp.solutions.hands
+mp_drawing = mp.solutions.drawing_utils
+hands = mp_hands.Hands(static_image_mode=True, max_num_hands=1, min_detection_confidence=0.5)
+
+# Define the ASL words
+asl_words = ["Hello", "Thanks", "Yes", "No", "Love", "Please"]  # Expand this list
+
+def extract_hand_landmarks(image):
+    """Extracts hand landmarks from an image."""
+    image_rgb = cv2.cvtColor(image, cv2.COLOR_BGR2RGB)
+    results = hands.process(image_rgb)
+    
+    if results.multi_hand_landmarks:
+        for hand_landmarks in results.multi_hand_landmarks:
+            landmarks = np.array([[lm.x, lm.y, lm.z] for lm in hand_landmarks.landmark]).flatten()
+            return landmarks
+    return None
+
+def classify_asl_word(image):
+    """Predicts the ASL word from a hand gesture."""
+    landmarks = extract_hand_landmarks(image)
+    if landmarks is None:
+        return "No hand detected"
+    
+    # Convert landmarks into a format suitable for the model
+    inputs = processor(images=image, return_tensors="pt")
+    
+    with torch.no_grad():
+        outputs = model(**inputs)
+    predicted_class_idx = outputs.logits.argmax(-1).item()
+    
+    return asl_words[predicted_class_idx] if predicted_class_idx < len(asl_words) else "Unknown sign"
+
+def process_video(video_path):
+    """Processes the uploaded video and returns detected ASL words."""
+    cap = cv2.VideoCapture(video_path)
+    detected_words = []
+
+    while cap.isOpened():
+        ret, frame = cap.read()
+        if not ret:
+            break
+        word = classify_asl_word(frame)
+        if word not in detected_words and word != "No hand detected":
+            detected_words.append(word)
+
+    cap.release()
+    return ", ".join(detected_words) if detected_words else "No ASL words detected"
+
+# Gradio Interface
+iface = gr.Interface(
+    fn=process_video,
+    inputs=gr.Video(type="file"),
+    outputs=gr.Textbox(label="Detected ASL Words"),
+    title="ASL Sign Language to Text",
+    description="Upload a video of ASL signs, and the model will translate them into text."
+)
+
+if __name__ == "__main__":
+    iface.launch()