update docs

examples/FasterRCNN/dataset/coco.py

@@ -162,7 +162,7 @@ class COCODetection(DatasetSplit):
             y2 = min(max(y2, 0), height)
             w, h = x2 - x1, y2 - y1
             # Require non-zero seg area and more than 1x1 box size
-            if obj['area'] > 1 and w > 0 and h > 0 and w * h >= 4:
+            if obj['area'] > 1 and w > 0 and h > 0:
                 all_boxes.append([x1, y1, x2, y2])
                 all_cls.append(self.COCO_id_to_category_id.get(obj['category_id'], obj['category_id']))
                 iscrowd = obj.get("iscrowd", 0)

examples/FasterRCNN/dataset/dataset.py

@@ -25,6 +25,7 @@ class DatasetSplit():
             Each list of numpy arrays corresponds to the mask for one instance.
             Each numpy array in the list is a polygon of shape Nx2,
             because one mask can be represented by N polygons.
+            Each row in the Nx2 array is a (x, y) coordinate.
 
             If your segmentation annotations are originally masks rather than polygons,
             either convert it, or the augmentation will need to be changed or skipped accordingly.