Update app.py

app.py

@@ -1,38 +1,43 @@
 import gradio as gr
-from transformers import DetrImageProcessor, DetrForObjectDetection
+from transformers import TrOCRProcessor, VisionEncoderDecoderModel
+from PIL import Image
 import torch
-from PIL import Image, ImageDraw
+from TTS.api import TTS
+import tempfile
 
-# Load pre-trained model and processor
-processor = DetrImageProcessor.from_pretrained("facebook/detr-resnet-50")
-model = DetrForObjectDetection.from_pretrained("facebook/detr-resnet-50")
+# Load OCR model
+processor = TrOCRProcessor.from_pretrained("microsoft/trocr-base-stage1")
+model = VisionEncoderDecoderModel.from_pretrained("microsoft/trocr-base-stage1")
 
-# Object detection function
-def detect_objects(image):
-    # Convert image and run model
-    inputs = processor(images=image, return_tensors="pt")
-    outputs = model(**inputs)
+# Load multilingual TTS model (supports Bangla)
+tts = TTS(model_name="tts_models/multilingual/multi-dataset/your_tts", progress_bar=False, gpu=False)
 
-    # Get outputs
-    target_sizes = torch.tensor([image.size[::-1]])  # PIL: (W, H) -> expected (H, W)
-    results = processor.post_process_object_detection(outputs, target_sizes=target_sizes, threshold=0.9)[0]
+def bangla_reader(image):
+    if image is None:
+        return "কোনো ছবি পাওয়া যায়নি।", None
 
-    # Draw boxes on the image
-    draw = ImageDraw.Draw(image)
-    for score, label, box in zip(results["scores"], results["labels"], results["boxes"]):
-        box = [round(i, 2) for i in box.tolist()]
-        draw.rectangle(box, outline="red", width=3)
-        draw.text((box[0], box[1]), f"{model.config.id2label[label.item()]}: {round(score.item(), 3)}", fill="red")
+    # OCR
+    pixel_values = processor(images=image, return_tensors="pt").pixel_values
+    generated_ids = model.generate(pixel_values)
+    ocr_text = processor.batch_decode(generated_ids, skip_special_tokens=True)[0]
 
-    return image
+    # TTS
+    with tempfile.NamedTemporaryFile(suffix=".wav", delete=False) as tmp:
+        tts.tts_to_file(text=ocr_text, file_path=tmp.name, language="bn", speaker="female")
+        audio_path = tmp.name
 
-# Launch Gradio interface
+    return f"OCR ফলাফল: {ocr_text}", audio_path
+
+# Gradio UI
 demo = gr.Interface(
-    fn=detect_objects,
-    inputs=gr.Image(source="camera", tool="editor", live=True),
-    outputs=gr.Image(type="pil"),
-    title="Real-Time Object Detection",
-    description="Open webcam and detect objects using facebook/detr-resnet-50"
+    fn=bangla_reader,
+    inputs=gr.Image(type="pil", label="বাংলা লেখা সম্বলিত ছবি দিন"),
+    outputs=[
+        gr.Textbox(label="OCR ফলাফল"),
+        gr.Audio(label="বাংলা কণ্ঠে পাঠ করুন")
+    ],
+    title="📖 বাংলা রিডার",
+    description="ছবির বাংলা লেখা পড়ে তা কণ্ঠে রূপান্তর করে শোনায়।"
 )
 
 if __name__ == "__main__":