def test_analyze_video(self):
base_path = os.path.join("..", "test_files", "video_indexing")
resource_path = os.path.join(base_path, "fic.02.mpg")
resource_id = "fic.02.mpg"
params = {c.VIDEO_INDEXING_PATH_KEY: base_path}
fe = VideoFaceExtractor(resource_path, resource_id, params)
fe.analyze_video()
def test_merge_consecutive_segments(self):
params = {c.MIN_SEGMENT_DURATION_KEY: 1}
fe = VideoFaceExtractor("path", "id", params)
segments = [
{
c.SEGMENT_START_KEY: 0,
c.SEGMENT_DURATION_KEY: 10000,
c.SEGMENT_COUNTER_KEY: 0,
c.FRAMES_KEY: [0, 1, 2, 3],
},
{c.SEGMENT_START_KEY: 10500, c.SEGMENT_DURATION_KEY: 4500, c.SEGMENT_COUNTER_KEY: 1, c.FRAMES_KEY: [4, 5]},
{c.SEGMENT_START_KEY: 15900, c.SEGMENT_DURATION_KEY: 14100, c.SEGMENT_COUNTER_KEY: 2, c.FRAMES_KEY: [6]},
{
c.SEGMENT_START_KEY: 32000,
c.SEGMENT_DURATION_KEY: 8000,
c.SEGMENT_COUNTER_KEY: 3,
c.FRAMES_KEY: [7, 8, 9],
},
{
c.SEGMENT_START_KEY: 45000,
c.SEGMENT_DURATION_KEY: 5000,
c.SEGMENT_COUNTER_KEY: 4,
c.FRAMES_KEY: [10, 11, 12, 13, 14, 15],
},
{
c.SEGMENT_START_KEY: 50000,
c.SEGMENT_DURATION_KEY: 5000,
c.SEGMENT_COUNTER_KEY: 5,
c.FRAMES_KEY: [16, 17],
},
]
fe.recognized_faces = [{c.ASSIGNED_LABEL_KEY: 1, c.SEGMENTS_KEY: segments}]
fe.merge_consecutive_segments()
segments = fe.recognized_faces[0][c.SEGMENTS_KEY]
self.assertEqual(len(segments), 3)
self.assertEqual(segments[0][c.SEGMENT_DURATION_KEY], 30000)
self.assertEqual(len(segments[0][c.FRAMES_KEY]), 7)
def test_delete_analysis_results(self):
self.test_analyze_video()
base_path = os.path.join("..", "test_files", "video_indexing")
resource_path = os.path.join(base_path, "fic.02.mpg")
resource_id = "fic.02"
params = {c.VIDEO_INDEXING_PATH_KEY: base_path}
fe = VideoFaceExtractor(resource_path, resource_id, params)
fe.delete_analysis_results()
exists = os.path.exists(fe.video_path)
self.assertFalse(exists)
def test_add_keyfaces_to_models(self):
self.test_analyze_video()
base_path = os.path.join("..", "test_files", "video_indexing")
resource_path = os.path.join(base_path, "fic.02.mpg")
resource_id = "fic.02"
params = {c.VIDEO_INDEXING_PATH_KEY: base_path}
fe = VideoFaceExtractor(resource_path, resource_id, params)
people = fe.get_people()
for person_dict in people:
label = person_dict[c.PERSON_COUNTER_KEY]
tag = str(label)
fe.add_keyface_to_models(label, label, tag)
def test_merge_labels(self):
fe = VideoFaceExtractor("path", "id")
segments_1 = ["1a", "1b", "1c"]
segments_2 = ["2a", "2b", "2c"]
segments_3 = ["3a", "3b", "3c"]
fe.recognized_faces = [
{c.ASSIGNED_LABEL_KEY: 1, c.SEGMENTS_KEY: segments_1},
{c.ASSIGNED_LABEL_KEY: 2, c.SEGMENTS_KEY: segments_2},
{c.ASSIGNED_LABEL_KEY: 1, c.SEGMENTS_KEY: segments_1},
{c.ASSIGNED_LABEL_KEY: 3, c.SEGMENTS_KEY: segments_3},
{c.ASSIGNED_LABEL_KEY: 1, c.SEGMENTS_KEY: segments_1},
{c.ASSIGNED_LABEL_KEY: 2, c.SEGMENTS_KEY: segments_2},
]
fe.merge_labels()
self.assertEqual(len(fe.recognized_faces), 3)
self.assertEqual(len(fe.recognized_faces[0][c.SEGMENTS_KEY]), 9)
def test_track_people_in_video(self):
base_path = os.path.join("..", "test_files", "video_indexing")
resource_path = os.path.join(base_path, "fic.02.mpg")
resource_id = "fic.02.mpg"
params = {c.VIDEO_INDEXING_PATH_KEY: base_path}
fe = VideoFaceExtractor(resource_path, resource_id, params)
fe.fps = 25.0
fe.cluster_faces_in_video()
fe.track_people_in_video()
fe.save_rec_people()