feat : Make predictions with different speed levels

Project/Backend/Server_Python/app.log

@@ -16,3 +16,18 @@
 2023-07-12 05:50:25,202 - INFO - Error. 'SignLanguagePredictionService' object has no attribute 'predict_sign_language_video2'
 2023-07-12 05:50:25,202 - INFO - Error. 'SignLanguagePredictionService' object has no attribute 'predict_sign_language_video2'
 2023-07-12 05:50:25,202 - INFO - Error. 'SignLanguagePredictionService' object has no attribute 'predict_sign_language_video2'
+2023-07-12 06:33:48,435 - INFO - Error. SignLanguagePredictionService.predict_sign_language_video_new2() missing 1 required positional argument: 'speed'
+2023-07-12 06:33:48,435 - INFO - Error. SignLanguagePredictionService.predict_sign_language_video_new2() missing 1 required positional argument: 'speed'
+2023-07-12 06:33:48,435 - INFO - Error. SignLanguagePredictionService.predict_sign_language_video_new2() missing 1 required positional argument: 'speed'
+2023-07-12 06:34:27,777 - INFO - Error. SignLanguagePredictionService.predict_sign_language_video_new2() missing 1 required positional argument: 'speed'
+2023-07-12 06:34:27,777 - INFO - Error. SignLanguagePredictionService.predict_sign_language_video_new2() missing 1 required positional argument: 'speed'
+2023-07-12 06:34:27,777 - INFO - Error. SignLanguagePredictionService.predict_sign_language_video_new2() missing 1 required positional argument: 'speed'
+2023-07-12 06:34:33,502 - INFO - Error. SignLanguagePredictionService.predict_sign_language_video_new2() missing 1 required positional argument: 'speed'
+2023-07-12 06:34:33,502 - INFO - Error. SignLanguagePredictionService.predict_sign_language_video_new2() missing 1 required positional argument: 'speed'
+2023-07-12 06:34:33,502 - INFO - Error. SignLanguagePredictionService.predict_sign_language_video_new2() missing 1 required positional argument: 'speed'
+2023-07-12 06:39:07,178 - INFO - Failed to make predictions. local variable 'threshold_percentage' referenced before assignment
+2023-07-12 06:39:07,178 - INFO - Failed to make predictions. local variable 'threshold_percentage' referenced before assignment
+2023-07-12 06:39:07,178 - INFO - Failed to make predictions. local variable 'threshold_percentage' referenced before assignment
+2023-07-12 06:39:07,180 - INFO - Error. 
+2023-07-12 06:39:07,180 - INFO - Error. 
+2023-07-12 06:39:07,180 - INFO - Error. 

Project/Backend/Server_Python/controllers/translate_controler.py

-from fastapi import APIRouter, File, HTTPException,UploadFile
+from fastapi import APIRouter, File, HTTPException, Query,UploadFile
 from pydantic import BaseModel
 import tensorflow as tf
 from core.logger import setup_logger
@@ -19,9 +19,10 @@ model = tf.keras.models.load_model('../ML_Models/sign_language_to_text/models/si
 CLASSES = mappings.classes
 NUM_CLASSES = len(mappings.classes)  # number of classes
 IMG_SIZE = 224  # image size
+speed_levels = mappings.speed_levels
 
 # Instantiate the service class
-prediction_service = SignLanguagePredictionService(model, CLASSES, mappings)
+prediction_service = SignLanguagePredictionService(model, CLASSES, mappings,speed_levels)
 
 
 @router.post("/upload/video", tags=["Sign Language"])
@@ -54,6 +55,17 @@ def predict_using_image(image_request: UploadFile = File(...)):
 def predict_using_video(video_request: UploadFile = File(...)):
     try:
         return prediction_service.predict_sign_language_video_new(video_request)
+    except Exception as e:
+        logger.info(f"Error. {e}")
+        raise HTTPException(
+            status_code=500,
+            detail="Request Failed."
+        )
+    
+@router.post('/predict-sign-language/video/speed_levels', tags=["Sign Language"])
+def predict_using_video(video_request: UploadFile = File(...), speed: int = Query(...)):
+    try:
+        return prediction_service.predict_sign_language_video_with_speed_levels(video_request, speed=speed)
     except Exception as e:
         logger.info(f"Error. {e}")
         raise HTTPException(

Project/Backend/Server_Python/services/translate_service.py

@@ -13,10 +13,11 @@ logger = setup_logger()
 IMG_SIZE = 224  # image size
 
 class SignLanguagePredictionService:
-    def __init__(self, model, classes, mappings):
+    def __init__(self, model, classes, mappings,speed_levels):
         self.model = model
         self.classes = classes
         self.mappings = mappings
+        self.speed_levels = speed_levels
 
     def predict_sign_language(self, image_request: UploadFile) -> Dict[str, str]:
         try:
@@ -149,6 +150,66 @@ class SignLanguagePredictionService:
                 status_code=500,
                 detail="Failed to make predictions"
             )
+        
+    def predict_sign_language_video_with_speed_levels(self, video_request: UploadFile, speed: int) -> Dict[str, str]:
+        try:
+            # Create a temporary file to save the video
+            video_location = f"files/{video_request.filename}"
+            with open(video_location, "wb") as file:
+                file.write(video_request.file.read())
+
+            # Read the video using OpenCV
+            video = cv2.VideoCapture(video_location)
+
+            predictions = []
+            final_predictions = []
+            frame_count = 0
+
+            # Determine the number of frames per sign based on the speed level
+            frames_per_sign = self.speed_levels.get(speed, 50)  # Default to level 1 if speed level is not provided
+
+            # Loop through the frames of the video
+            while frame_count < 50:
+                success, frame = video.read()
+                if not success:
+                    break
+
+                # Preprocess the frame
+                frame = cv2.cvtColor(frame, cv2.COLOR_BGR2RGB)
+                frame = cv2.resize(frame, (IMG_SIZE, IMG_SIZE))
+                frame = extract_hand_shape(frame)
+                frame = np.array([frame], dtype=np.float32) / 255.0
+
+                # Make prediction
+                prediction = self.model.predict(frame)
+                class_index = np.argmax(prediction)
+                class_name = self.classes[class_index]
+
+                sinhala_letter = self.mappings.letter_mapping.get(class_name, 'Unknown')
+
+                # Store the prediction for the frame
+                predictions.append(sinhala_letter)
+                frame_count += 1
+                threshold_percentage = 60
+
+                # Check if the required number of frames per sign has been reached
+                if frame_count % frames_per_sign == 0:
+                    predictions = get_predicted_percentage(predictions, threshold_percentage)
+                    final_predictions = final_predictions+predictions
+                    predictions = []
+
+            video.release()
+
+            # Delete the video file
+            os.remove(video_location)
+
+            return {'frame_count': frame_count, 'predictions': final_predictions}
+        except Exception as e:
+            logger.info(f"Failed to make predictions. {e}")
+            raise HTTPException(
+                status_code=500,
+                detail="Failed to make predictions"
+            )
 
 def extract_hand_shape(image):
     gray = cv2.cvtColor(image, cv2.COLOR_RGB2GRAY)

Project/Backend/Server_Python/utils/mappings.py

@@ -27,4 +27,13 @@ classes =['A',
  'Ohh',
  'T',
  'Uh',
- 'Uhh']
\ No newline at end of file
+ 'Uhh']
+
+
+speed_levels = {
+    1: 50,  # 10 frames per sign for level 1
+    2: 40,  # 20 frames per sign for level 2
+    3: 30,  # 30 frames per sign for level 3
+    4: 20,  # 40 frames per sign for level 4
+    5: 10   # 50 frames per sign for level 5
+}
\ No newline at end of file