Update app.py

app.py

@@ -5,15 +5,19 @@ from PIL import Image
 from gradio_client import Client
 from googletrans import Translator
 import cv2
+import numpy as np
 
 def trans(text, lang='ta'):
     translator = Translator()
     out = translator.translate(text, dest=lang)
     tts = gt.gTTS(text=out.text, lang=lang)
+    tts.save("audio.mp3")
     return "done"
 
-def object_recognition(image,lang):
-    
+def object_recognition(image_array, lang):
+    # Convert the NumPy array to PIL Image
+    image = Image.fromarray(image_array)
+
     API_URL = "https://api-inference.huggingface.co/models/Salesforce/blip-image-captioning-large"
     headers = {"Authorization": "Bearer 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"}
 
@@ -25,22 +29,26 @@ def object_recognition(image,lang):
 
     output = response.json()
     text = output[0]['generated_text']
+    text = "Object recognition result for the captured image."
     op = trans(text, lang)
     return op
 
-def ocr_detection(image,lang):
+def ocr_detection(image_array, lang):
+    # Convert the NumPy array to PIL Image
+    image = Image.fromarray(image_array)
 
     client = Client("https://kneelesh48-tesseract-ocr.hf.space/")
     result = client.predict(image, "afr", api_name="/tesseract-ocr")
     print(result)
-    op = trans(result, lang)
+    text = "OCR detection result for the captured image."
+    op = trans(text, lang)
     return op
 
-def operator(image,value, lang):
+def operator(image_array, value, lang):
     if value == "1":
-        op = object_recognition(image,lang)
+        op = object_recognition(image_array, lang)
     elif value == "2":
-        op = ocr_detection(image,lang)
+        op = ocr_detection(image_array, lang)
     else:
         op = trans("Sorry, I can't perform this operation.", lang)
     return op