data_juicer.utils.ASD_mapper_utils module#

data_juicer.utils.ASD_mapper_utils.scene_detect(videoFilePath)[source]#

data_juicer.utils.ASD_mapper_utils.inference_video(video_array, DET)[source]#

data_juicer.utils.ASD_mapper_utils.get_video_array_cv2(videoFilePath)[source]#

data_juicer.utils.ASD_mapper_utils.bb_intersection_over_union(boxA, boxB, evalCol=False)[source]#

data_juicer.utils.ASD_mapper_utils.track_shot(sceneFaces, numFailedDet=8, minTrack=10)[source]#

data_juicer.utils.ASD_mapper_utils.find_human_bounding_box(face_bbox, human_bboxes)[source]#

data_juicer.utils.ASD_mapper_utils.update_negative_ones(values)[source]#

data_juicer.utils.ASD_mapper_utils.detect_and_mark_anomalies(data, window_size=7, std_multiplier=2)[source]#

data_juicer.utils.ASD_mapper_utils.crop_video_with_facetrack(video_array, track, cropFile, audioFilePath, is_empty=False)[source]#

data_juicer.utils.ASD_mapper_utils.evaluate_network(files, s, pycropPath)[source]#

data_juicer.utils.ASD_mapper_utils.visualization(tracks, scores, video_array, pyaviPath)[source]#

data_juicer.utils.ASD_mapper_utils.calculate_good_matches(matches, ratio=0.75)[source]#

data_juicer.utils.ASD_mapper_utils.find_max_intersection_and_remaining_dicts(dicts)[source]#

data_juicer.utils.ASD_mapper_utils.get_faces_array(frame, s, x, y)[source]#

data_juicer.utils.ASD_mapper_utils.order_track_distance(track1, track2, video_array)[source]#

data_juicer.utils.ASD_mapper_utils.update_remain(remaining_dicts, pop_item)[source]#

data_juicer.utils.ASD_mapper_utils.order_merge_tracks(track1, track2)[source]#

data_juicer.utils.ASD_mapper_utils.post_merge(vidTracks, video_array)[source]#

data_juicer.utils.ASD_mapper_utils.longest_continuous_actives(arr)[source]#

data_juicer.utils.ASD_mapper_utils.annotate_video_with_bounding_boxes_with_audio(video_path, q_human_video_track_bbox, output_path)[source]#

data_juicer.utils.ASD_mapper_utils.annotate_video_with_bounding_boxes_withText_with_audio(video_path, q_human_video_track_bbox, output_path, numbers)[source]#

data_juicer.utils.ASD_mapper_utils.annotate_video_with_bounding_boxes(video_array, frame_indices, bounding_boxes, output_path)[source]#

Annotates specified frames in the video with bounding boxes and saves the result to a new video file.

Parameters:

video_array – Input video as a numpy array with shape (num_frames, height, width, channels).
frame_indices – List of frame indices to annotate.
bounding_boxes – Array of bounding box coordinates with shape (num_frames_to_annotate, 4), where each bounding box is (x, y, w, h).
output_path – Path to save the output video.

data_juicer.utils.ASD_mapper_utils.crop_from_array(frame_before_crop, coords)[source]#