Update app.py

app.py

@@ -2,44 +2,66 @@ import gradio as gr
 import cv2
 import numpy as np
 import easyocr
+import tempfile
 
-# AI model load ho raha hai jo English text padhega
+# AI model load ho raha hai
 reader = easyocr.Reader(['en'], gpu=False)
 
-def remove_text_from_image(img):
-    # Photo ko array me convert karna
-    img = np.array(img)
+def remove_text_from_video(video_path):
+    if video_path is None:
+        return None
+        
+    cap = cv2.VideoCapture(video_path)
+    fps = int(cap.get(cv2.CAP_PROP_FPS))
+    width = int(cap.get(cv2.CAP_PROP_FRAME_WIDTH))
+    height = int(cap.get(cv2.CAP_PROP_FRAME_HEIGHT))
     
-    # Photo me text kahan hai, wo dhundhna
-    results = reader.readtext(img)
+    # Processed video save karne ke liye temporary file
+    temp_file = tempfile.NamedTemporaryFile(delete=False, suffix='.mp4')
+    output_path = temp_file.name
     
-    # Ek khali (black) mask banana
-    mask = np.zeros(img.shape[:2], dtype=np.uint8)
+    fourcc = cv2.VideoWriter_fourcc(*'mp4v')
+    out = cv2.VideoWriter(output_path, fourcc, fps, (width, height))
     
-    # Jahan text mila hai, wahan mask par safed (white) box banana
-    for (bbox, text, prob) in results:
-        (tl, tr, br, bl) = bbox
-        tl = (int(tl[0]), int(tl[1]))
-        br = (int(br[0]), int(br[1]))
-        cv2.rectangle(mask, tl, br, 255, -1)
-        
-    # Text ke kinaro ko thoda bada karna taaki pura text cover ho jaye
-    kernel = np.ones((5,5), np.uint8)
-    mask = cv2.dilate(mask, kernel, iterations=1)
+    # Hugging Face free tier ko crash hone se bachane ke liye 150 frames (approx 5 sec) ki limit
+    max_frames = 150 
+    frame_count = 0
     
-    # Asli magic: Text ko hata kar background fill karna (Inpainting)
-    result = cv2.inpaint(img, mask, inpaintRadius=3, flags=cv2.INPAINT_TELEA)
+    while cap.isOpened():
+        ret, frame = cap.read()
+        if not ret or frame_count >= max_frames:
+            break
+            
+        # Text dhundhna aur mask banana
+        results = reader.readtext(frame)
+        mask = np.zeros(frame.shape[:2], dtype=np.uint8)
+        
+        for (bbox, text, prob) in results:
+            (tl, tr, br, bl) = bbox
+            cv2.rectangle(mask, (int(tl[0]), int(tl[1])), (int(br[0]), int(br[1])), 255, -1)
+            
+        # Mask ko thoda bada karna taaki text puri tarah cover ho
+        kernel = np.ones((5,5), np.uint8)
+        mask = cv2.dilate(mask, kernel, iterations=1)
+        
+        # Frame se text hata kar background fill karna (Inpainting)
+        result_frame = cv2.inpaint(frame, mask, 3, cv2.INPAINT_TELEA)
+        
+        out.write(result_frame)
+        frame_count += 1
+        
+    cap.release()
+    out.release()
     
-    return result
+    return output_path
 
-# Gradio ka User Interface (UI) banana
+# Gradio Video UI
 interface = gr.Interface(
-    fn=remove_text_from_image,
-    inputs=gr.Image(type="numpy", label="Upload Image (Text wali photo daalein)"),
-    outputs=gr.Image(type="numpy", label="Result (Bina text wali photo)"),
-    title="Hyco AI Text Remover",
-    description="Is tool me photo upload karein, aur AI automatically usme se captions aur text hata dega."
+    fn=remove_text_from_video,
+    inputs=gr.Video(label="Upload Video (Test karne ke liye 2-5 sec ki clip daalein)"),
+    outputs=gr.Video(label="Result Video"),
+    title="Hyco Video Text Remover",
+    description="Isme apni clip upload karein. AI har frame se captions dhundhega aur unhe automatically hata dega."
 )
 
-# App ko start karna
 interface.launch()