adding some more description

crabs/tracker/evaluate_tracker.py

@@ -170,11 +170,11 @@ def evaluate_mota(
         Parameters
         ----------
         gt_boxes : np.ndarray
-            Ground truth bounding boxes of objects.
+            Ground truth bounding boxes of objects with shape of (N, 4) with (x1, y1, x2, y2).
         gt_ids : np.ndarray
-            Ground truth IDs corresponding to the bounding boxes.
+            Ground truth IDs corresponding to the bounding boxes with shape of (N, 1).
         tracked_boxes : np.ndarray
-            Tracked bounding boxes of objects.
+            Tracked bounding boxes of objects with shape of (N, 5) with (x1, y1, x2, y2, id).
         iou_threshold : float
             Intersection over Union (IoU) threshold for considering a match.
         prev_frame_id_map : Optional[Dict[int, int]]

crabs/tracker/track_video.py

@@ -143,7 +143,7 @@ def update_tracking(self, prediction: dict) -> list[list[float]]:
         """
         pred_sort = prep_sort(prediction, self.config["score_threshold"])
         tracked_boxes = self.sort_tracker.update(pred_sort)
-        self.tracked_list.append(tracked_boxes)
+        self.tracked_bbox_id.append(tracked_boxes)
         return tracked_boxes
 
     def run_tracking(self):
@@ -160,7 +160,7 @@ def run_tracking(self):
         # In any case run inference
         # initialisation
         frame_idx = 0
-        self.tracked_list = []
+        self.tracked_bbox_id = []
 
         # Loop through frames of the video in batches
         while self.video.isOpened():
@@ -199,7 +199,7 @@ def run_tracking(self):
         if self.args.gt_path:
             evaluation = TrackerEvaluate(
                 self.args.gt_path,
-                self.tracked_list,
+                self.tracked_bbox_id,
                 self.config["iou_threshold"],
             )
             evaluation.run_evaluation()