Update app.py

app.py

@@ -1,78 +1,53 @@
+import streamlit as st
 import cv2
-from PIL import Image
 import numpy as np
-import pytesseract
-from googletrans import Translator
-
-def overlay_translation(image_path, src_lang='hi', dest_lang='en'):
-    # Load image
-    image = cv2.imread(image_path)
-    image_rgb = cv2.cvtColor(image, cv2.COLOR_BGR2RGB)
-
-    # OCR data with boxes
-    data = pytesseract.image_to_data(image_rgb, lang=src_lang, output_type=pytesseract.Output.DICT)
-
-    translator = Translator()
-    translated_image = image_rgb.copy()
-
-    for i in range(len(data['text'])):
-        text = data['text'][i].strip()
-        if text != "":
-            x, y, w, h = data['left'][i], data['top'][i], data['width'][i], data['height'][i]
-
-            try:
-                translated_text = translator.translate(text, src=src_lang, dest=dest_lang).text
-            except:
-                translated_text = text  # fallback
-
-            # White rectangle behind text for better visibility
-            cv2.rectangle(translated_image, (x, y), (x+w, y+h), (255,255,255), -1)
-
-            # Add translated text (in black, centered)
-            font_scale = h / 35
-            font_thickness = 2
-            cv2.putText(translated_image, translated_text, (x, y + h - 5), 
-                        cv2.FONT_HERSHEY_SIMPLEX, font_scale, (0,0,0), font_thickness, cv2.LINE_AA)
-
-    # Save or return the translated image
-    translated_image_pil = Image.fromarray(translated_image)
-    return translated_image_pil
-import cv2
 from PIL import Image
-import numpy as np
 import pytesseract
-from googletrans import Translator
+from deep_translator import GoogleTranslator
 
-def overlay_translation(image_path, src_lang='hi', dest_lang='en'):
-    # Load image
-    image = cv2.imread(image_path)
-    image_rgb = cv2.cvtColor(image, cv2.COLOR_BGR2RGB)
+st.set_page_config(page_title="Image Translator App", layout="centered")
+st.title("📄 Image Translator")
+st.write("Upload an image with text and get it translated directly on the image.")
 
-    # OCR data with boxes
-    data = pytesseract.image_to_data(image_rgb, lang=src_lang, output_type=pytesseract.Output.DICT)
+uploaded_file = st.file_uploader("Choose an image...", type=["jpg", "jpeg", "png"])
 
-    translator = Translator()
-    translated_image = image_rgb.copy()
+if uploaded_file is not None:
+    image = Image.open(uploaded_file)
+    image_np = np.array(image)
 
-    for i in range(len(data['text'])):
-        text = data['text'][i].strip()
-        if text != "":
-            x, y, w, h = data['left'][i], data['top'][i], data['width'][i], data['height'][i]
+    st.image(image, caption="Original Image", use_column_width=True)
 
-            try:
-                translated_text = translator.translate(text, src=src_lang, dest=dest_lang).text
-            except:
-                translated_text = text  # fallback
+    # Convert image to RGB (OpenCV uses BGR by default)
+    image_rgb = cv2.cvtColor(image_np, cv2.COLOR_RGB2BGR)
 
-            # White rectangle behind text for better visibility
-            cv2.rectangle(translated_image, (x, y), (x+w, y+h), (255,255,255), -1)
+    # OCR - get data with bounding boxes
+    data = pytesseract.image_to_data(image_rgb, output_type=pytesseract.Output.DICT)
 
-            # Add translated text (in black, centered)
-            font_scale = h / 35
-            font_thickness = 2
-            cv2.putText(translated_image, translated_text, (x, y + h - 5), 
-                        cv2.FONT_HERSHEY_SIMPLEX, font_scale, (0,0,0), font_thickness, cv2.LINE_AA)
+    translator = GoogleTranslator(source='auto', target='hi')
 
-    # Save or return the translated image
-    translated_image_pil = Image.fromarray(translated_image)
-    return translated_image_pil
+    n_boxes = len(data['text'])
+    for i in range(n_boxes):
+        if int(data['conf'][i]) > 60:  # confidence threshold
+            x, y, w, h = data['left'][i], data['top'][i], data['width'][i], data['height'][i]
+            text = data['text'][i]
+            if text.strip() != "":
+                try:
+                    translated_text = translator.translate(text)
+
+                    # Draw white rectangle as background for text
+                    cv2.rectangle(image_rgb, (x, y), (x + w, y + h), (255, 255, 255), -1)
+
+                    # Adjust font scale based on box height
+                    font_scale = h / 35
+                    font_thickness = 1
+
+                    # Overlay translated text, vertically aligned inside the box
+                    cv2.putText(image_rgb, translated_text, (x, y + h - 5),
+                                cv2.FONT_HERSHEY_SIMPLEX, font_scale, (0, 0, 0),
+                                font_thickness, cv2.LINE_AA)
+                except Exception as e:
+                    print(f"Translation error for '{text}': {e}")
+
+    # Convert back to RGB for PIL display
+    result_image = cv2.cvtColor(image_rgb, cv2.COLOR_BGR2RGB)
+    st.image(result_image, caption="Translated Image", use_column_width=True)