Update video_processing.py

video_processing.py

@@ -10,8 +10,7 @@ from face_analysis import get_face_embedding, cluster_faces, organize_faces_by_p
 from pose_analysis import calculate_posture_score, draw_pose_landmarks
 from anomaly_detection import anomaly_detection
 from visualization import plot_mse, plot_mse_histogram, plot_mse_heatmap
-from utils import frame_to_timecode, parse_transcription, get_sentences_before_anomalies
-from transcribe import transcribe
+from utils import frame_to_timecode
 import pandas as pd
 from facenet_pytorch import MTCNN
 import torch
@@ -183,23 +182,6 @@ def process_video(video_path, anomaly_threshold, desired_fps, progress=None):
             traceback.print_exc()
             return (f"Error in video processing: {str(e)}",) + (None,) * 14
 
-        # Add transcription
-        progress(0.96, "Transcribing video")
-        transcription_output = transcribe(video_path, transcribe_to_text=True, transcribe_to_srt=False,
-                                          target_language='en')
-
-        # Parse the transcription output to get sentences and their timecodes
-        sentences_with_timecodes = parse_transcription(transcription_output, video_duration)
-
-        # Get anomaly timecodes
-        anomaly_timecodes_features = [df[df['Frame'] == frame]['Timecode'].iloc[0] for frame in
-                                      anomaly_frames_embeddings]
-        anomaly_timecodes_posture = [df[df['Frame'] == frame]['Timecode'].iloc[0] for frame in anomaly_frames_posture]
-
-        anomaly_sentences_features = get_sentences_before_anomalies(sentences_with_timecodes,
-                                                                    anomaly_timecodes_features)
-        anomaly_sentences_posture = get_sentences_before_anomalies(sentences_with_timecodes,
-                                                                   anomaly_timecodes_posture)
         progress(1.0, "Preparing results")
         results = f"Number of persons detected: {num_clusters}\n\n"
         results += "Breakdown:\n"
@@ -259,9 +241,7 @@ def process_video(video_path, anomaly_threshold, desired_fps, progress=None):
             anomaly_faces_embeddings,
             anomaly_frames_posture_images,
             aligned_faces_folder,
-            frames_folder,
-            anomaly_sentences_features,
-            anomaly_sentences_posture
+            frames_folder
         )
 
 def is_frontal_face(landmarks, threshold=40):