Update app.py

app.py

@@ -1,4 +1,4 @@
-# app.py (Enhanced to display both image and text)
+# app.py
 
 import gradio as gr
 import torch
@@ -30,32 +30,41 @@ import matplotlib.pyplot as plt
 ethiopic_font = fm.FontProperties(fname=FONT_PATH, size=15)
 pil_font = ImageFont.truetype(FONT_PATH, size=20)
 
-def recognize_and_overlay(image: Image.Image):
+def recognize_text(image: Image.Image) -> str:
+    """
+    Function to recognize text from an image.
+    """
+    # Preprocess the image
+    input_tensor = preprocess_image(image).unsqueeze(0).to(device)  # [1, 3, 224, 224]
+    
+    # Perform inference
+    with torch.no_grad():
+        log_probs = model(input_tensor)  # [H*W, 1, vocab_size]
+    
+    # Decode predictions
+    recognized_texts = decode_predictions(log_probs)
+    
+    return recognized_texts[0]
+
+def recognize_and_overlay(image: Image.Image) -> Image.Image:
     """
     Function to recognize text and overlay it on the image.
-    Returns both the modified image and the recognized text.
     """
     recognized_text = recognize_text(image)
     
-    # Create a copy of the image to draw on
-    image_with_text = image.copy()
-    
     # Overlay text on the image
-    draw = ImageDraw.Draw(image_with_text)
+    draw = ImageDraw.Draw(image)
     text_position = (10, 10)  # Top-left corner
     text_color = (255, 0, 0)  # Red color
     draw.text(text_position, f"Recognized: {recognized_text}", font=pil_font, fill=text_color)
     
-    return image_with_text, recognized_text
+    return image
 
 # Define Gradio Interface
 iface = gr.Interface(
     fn=recognize_and_overlay,
     inputs=gr.Image(type="pil", label="Upload Image"),
-    outputs=[
-        gr.Image(type="pil", label="Image with Recognized Text"),
-        gr.Textbox(label="Recognized Text")
-    ],
+    outputs=gr.Image(type="pil", label="Image with Recognized Text"),
     title="Amharic Text Recognition",
     description="Upload an image containing Amharic text. The app will recognize and overlay the text on the image."
 )