perform training on negative samples (images with no annotation)

detectron/core/config.py

@@ -133,6 +133,9 @@
 # faster)
 __C.TRAIN.ASPECT_GROUPING = True
 
+# Include background images (images with no annotation) in the training
+__C.TRAIN.INCLUDE_BKG_IMAGES = False
+
 # ---------------------------------------------------------------------------- #
 # RPN training options
 # ---------------------------------------------------------------------------- #

detectron/datasets/roidb.py

@@ -115,6 +115,9 @@ def is_valid(entry):
         # Valid images have:
         #   (1) At least one foreground RoI OR
         #   (2) At least one background RoI
+        if cfg.TRAIN.INCLUDE_BKG_IMAGES and len(entry['boxes']) == 0:
+            return True  # this is a bkg image
+
         overlaps = entry['max_overlaps']
         # find boxes with sufficient overlap
         fg_inds = np.where(overlaps >= cfg.TRAIN.FG_THRESH)[0]

detectron/roi_data/rpn.py

@@ -199,7 +199,13 @@ def _get_rpn_blobs(im_height, im_width, foas, all_anchors, gt_boxes):
     # (samples with replacement, but since the set of bg inds is large most
     # samples will not have repeats)
     num_bg = cfg.TRAIN.RPN_BATCH_SIZE_PER_IM - np.sum(labels == 1)
-    bg_inds = np.where(anchor_to_gt_max < cfg.TRAIN.RPN_NEGATIVE_OVERLAP)[0]
+    if len(gt_boxes) > 0:
+        bg_inds = np.where(anchor_to_gt_max < cfg.TRAIN.RPN_NEGATIVE_OVERLAP)[0]
+    else:
+        # any idx can be considered as negative example
+        assert cfg.TRAIN.INCLUDE_BKG_IMAGES
+        bg_inds = np.array(range(len(labels)))
+
     if len(bg_inds) > num_bg:
         enable_inds = bg_inds[npr.randint(len(bg_inds), size=num_bg)]
     else:
@@ -209,9 +215,10 @@ def _get_rpn_blobs(im_height, im_width, foas, all_anchors, gt_boxes):
     bg_inds = np.where(labels == 0)[0]
 
     bbox_targets = np.zeros((num_inside, 4), dtype=np.float32)
-    bbox_targets[fg_inds, :] = data_utils.compute_targets(
-        anchors[fg_inds, :], gt_boxes[anchor_to_gt_argmax[fg_inds], :]
-    )
+    if len(gt_boxes) > 0:
+        bbox_targets[fg_inds, :] = data_utils.compute_targets(
+            anchors[fg_inds, :], gt_boxes[anchor_to_gt_argmax[fg_inds], :]
+        )
 
     # Bbox regression loss has the form:
     #   loss(x) = weight_outside * L(weight_inside * x)