Update in Centroid Trackers

README.md

@@ -26,7 +26,7 @@ detector.YOLOv3
 `YOLOv3 + CentroidTracker` |  `TF-MobileNetSSD + CentroidTracker`
 :-------------------------:|:-------------------------:
 ![Cars with YOLO][cars-yolo-output]  |  ![Cows with tf-SSD][cows-tf-ssd-output]
-Video source: [link](https://flic.kr/p/89KYXt) | Video source: [link](https://flic.kr/p/26WeEWy)
+Video source: [link](https://flic.kr/p/L6qyxj) | Video source: [link](https://flic.kr/p/26WeEWy)
 
 
 ### Installation

examples/example_scripts/mot_TF_SSDMobileNet.py

@@ -1,6 +1,6 @@
 import cv2 as cv
 from motrackers.detectors import TF_SSDMobileNetV2
-from motrackers import CentroidTracker, CentroidKF_Tracker
+from motrackers import CentroidTracker, CentroidKF_Tracker, SORT
 from motrackers.utils import draw_tracks
 
 
@@ -9,7 +9,7 @@ def main(video_path, weights_path, config_path, use_gpu, tracker):
     model = TF_SSDMobileNetV2(
         weights_path=weights_path,
         configfile_path=config_path,
-        confidence_threshold=0.3,
+        confidence_threshold=0.4,
         nms_threshold=0.2,
         draw_bboxes=True,
         use_gpu=use_gpu
@@ -24,15 +24,13 @@ def main(video_path, weights_path, config_path, use_gpu, tracker):
             break
 
         bboxes, confidences, class_ids = model.detect(image)
-
         tracks = tracker.update(bboxes, confidences, class_ids)
-
-        updated_image = model.draw_bboxes(image, bboxes, confidences, class_ids)
+        updated_image = model.draw_bboxes(image.copy(), bboxes, confidences, class_ids)
 
         updated_image = draw_tracks(updated_image, tracks)
 
         cv.imshow("image", updated_image)
-        if cv.waitKey(1) & 0xFF == ord('q'):
+        if cv.waitKey(0) & 0xFF == ord('q'):
             break
 
     cap.release()
@@ -55,7 +53,8 @@ def main(video_path, weights_path, config_path, use_gpu, tracker):
     )
 
     parser.add_argument(
-        '--config', '-c', type=str,
+        '--config', '-c',
+        type=str,
         default="./../pretrained_models/tensorflow_weights/ssd_mobilenet_v2_coco_2018_03_29.pbtxt",
         help='path to config file of Caffe-MobileNetSSD (`.pbtxt` file).'
     )
@@ -66,15 +65,17 @@ def main(video_path, weights_path, config_path, use_gpu, tracker):
     )
 
     parser.add_argument(
-        '--tracker', type=str, default='CentroidKF_Tracker',
-        help="Tracker used to track objects. Options include ['CentroidTracker', 'CentroidKF_Tracker']")
+        '--tracker', type=str, default='SORT',
+        help="Tracker used to track objects. Options include ['CentroidTracker', 'CentroidKF_Tracker', 'SORT']")
 
     args = parser.parse_args()
 
     if args.tracker == 'CentroidTracker':
-        tracker = CentroidTracker(max_lost=5, tracker_output_format='mot_challenge')
+        tracker = CentroidTracker(max_lost=0, tracker_output_format='mot_challenge')
     elif args.tracker == 'CentroidKF_Tracker':
-        tracker = CentroidKF_Tracker(max_lost=3, tracker_output_format='mot_challenge')
+        tracker = CentroidKF_Tracker(max_lost=0, tracker_output_format='mot_challenge')
+    elif args.tracker == 'SORT':
+        tracker = SORT(max_lost=2, tracker_output_format='mot_challenge', iou_threshold=0.5, time_step=1)
     else:
         raise NotImplementedError
 

motrackers/__init__.py

@@ -8,6 +8,6 @@
 
 
 from motrackers.tracker import Tracker as CentroidTracker
-from motrackers.tracker import CentroidKF_Tracker
-from motrackers.iou_tracker import IOUTracker
+from motrackers.centroid_kf_tracker import CentroidKF_Tracker
 from motrackers.sort_tracker import SORT
+# from motrackers.iou_tracker import IOUTracker

motrackers/centroid_kf_tracker.py

@@ -0,0 +1,151 @@
+from collections import OrderedDict
+import numpy as np
+from scipy.spatial import distance
+from scipy.optimize import linear_sum_assignment
+from motrackers.tracker import Tracker
+from motrackers.track import KFTrackCentroid
+from motrackers.utils.misc import get_centroid
+
+
+def assign_tracks2detection_centroid_distances(bbox_tracks, bbox_detections, distance_threshold=10.):
+    """
+    Assigns detected bounding boxes to tracked bounding boxes using IoU as a distance metric.
+
+    Args:
+        bbox_tracks (numpy.ndarray): tracked bounding boxes with shape (n, 4) each row as (xmin, ymin, width, height).
+        bbox_detections (numpy.ndarray): detection bounding boxes with shape (m, 4) each row as (xmin, ymin, width, height).
+        distance_threshold (float): Minimum distance between the tracked object and new detection to consider for assignment.
+
+    Returns:
+        tuple: Tuple containing the following elements:
+            - matches (numpy.ndarray):  Array of shape `(n, 2)` where `n` is number of pairs formed after
+                matching tracks to detections. This is an array of tuples with each element as matched pair
+                of indices`(track_index, detection_index)`.
+            - unmatched_detections (numpy.ndarray):  Array of shape `(m,)` where `m` is number of unmatched detections.
+            - unmatched_tracks (numpy.ndarray):  Array of shape `(k,)` where `k` is the number of unmatched tracks.
+    """
+
+    if (bbox_tracks.size == 0) or (bbox_detections.size == 0):
+        return np.empty((0, 2), dtype=int), np.arange(len(bbox_detections), dtype=int), np.empty((0,), dtype=int)
+
+    if len(bbox_tracks.shape) == 1:
+        bbox_tracks = bbox_tracks[None, :]
+
+    if len(bbox_detections.shape) == 1:
+        bbox_detections = bbox_detections[None, :]
+
+    estimated_track_centroids = get_centroid(bbox_tracks)
+    detection_centroids = get_centroid(bbox_detections)
+    centroid_distances = distance.cdist(estimated_track_centroids, detection_centroids)
+
+    assigned_tracks, assigned_detections = linear_sum_assignment(centroid_distances)
+
+    unmatched_detections, unmatched_tracks = [], []
+
+    for d in range(bbox_detections.shape[0]):
+        if d not in assigned_detections:
+            unmatched_detections.append(d)
+
+    for t in range(bbox_tracks.shape[0]):
+        if t not in assigned_tracks:
+            unmatched_tracks.append(t)
+
+    # filter out matched with high distance between centroids
+    matches = []
+    for t, d in zip(assigned_tracks, assigned_detections):
+        if centroid_distances[t, d] > distance_threshold:
+            unmatched_detections.append(d)
+            unmatched_tracks.append(t)
+        else:
+            matches.append((t, d))
+
+    if len(matches):
+        matches = np.array(matches)
+    else:
+        matches = np.empty((0, 2), dtype=int)
+
+    return matches, np.array(unmatched_detections), np.array(unmatched_tracks)
+
+
+class CentroidKF_Tracker(Tracker):
+    """
+    Kalman filter based tracking of multiple detected objects.
+
+    Parameters
+    ----------
+    max_lost : int
+        Maximum number of consecutive frames object was not detected.
+    tracker_output_format : str
+        Output format of the tracker.
+    process_noise_scale : float or numpy.ndarray
+        Process noise covariance matrix of shape (3, 3) or covariance magnitude as scalar value.
+    measurement_noise_scale : float or numpy.ndarray
+        Measurement noise covariance matrix of shape (1,) or covariance magnitude as scalar value.
+    time_step : int or float
+        Time step for Kalman Filter.
+    """
+
+    def __init__(
+            self,
+            max_lost=1,
+            centroid_distance_threshold=30.,
+            tracker_output_format='mot_challenge',
+            process_noise_scale=1.0,
+            measurement_noise_scale=1.0,
+            time_step=1
+    ):
+        self.time_step = time_step
+        self.process_noise_scale = process_noise_scale
+        self.measurement_noise_scale = measurement_noise_scale
+        self.centroid_distance_threshold = centroid_distance_threshold
+        self.kalman_trackers = OrderedDict()
+        super().__init__(max_lost, tracker_output_format)
+
+    def _add_track(self, frame_id, bbox, detection_confidence, class_id, **kwargs):
+        self.tracks[self.next_track_id] = KFTrackCentroid(
+            self.next_track_id, frame_id, bbox, detection_confidence, class_id=class_id,
+            data_output_format=self.tracker_output_format, process_noise_scale=self.process_noise_scale,
+            measurement_noise_scale=self.measurement_noise_scale, **kwargs
+        )
+        self.next_track_id += 1
+
+    def update(self, bboxes, detection_scores, class_ids):
+        self.frame_count += 1
+        bbox_detections = np.array(bboxes, dtype='int')
+
+        track_ids = list(self.tracks.keys())
+        bbox_tracks = []
+        for track_id in track_ids:
+            bbox_tracks.append(self.tracks[track_id].predict())
+        bbox_tracks = np.array(bbox_tracks)
+
+        matches, unmatched_detections, unmatched_tracks = assign_tracks2detection_centroid_distances(
+            bbox_tracks, bbox_detections, distance_threshold=self.centroid_distance_threshold
+        )
+
+        for i in range(matches.shape[0]):
+            t, d = matches[i, :]
+            track_id = track_ids[t]
+            bbox = bboxes[d, :]
+            cid = class_ids[d]
+            confidence = detection_scores[d]
+            self._update_track(track_id, self.frame_count, bbox, confidence, cid, lost=0)
+
+        for d in unmatched_detections:
+            bbox = bboxes[d, :]
+            cid = class_ids[d]
+            confidence = detection_scores[d]
+            self._add_track(self.frame_count, bbox, confidence, cid)
+
+        for t in unmatched_tracks:
+            track_id = track_ids[t]
+            bbox = bbox_tracks[t, :]
+            confidence = self.tracks[track_id].detection_confidence
+            cid = self.tracks[track_id].class_id
+            self._update_track(track_id, self.frame_count, bbox, confidence, cid, lost=1)
+
+            if self.tracks[track_id].lost > self.max_lost:
+                self._remove_track(track_id)
+
+        outputs = self._get_tracks(self.tracks)
+        return outputs

motrackers/iou_tracker.py

@@ -26,7 +26,7 @@ class IOUTracker(Tracker):
 
     def __init__(
             self,
-            max_lost=2,
+            max_lost=0,
             iou_threshold=0.5,
             min_detection_confidence=0.4,
             max_detection_confidence=0.7,