def compute_targets(proposals, gt_rois, iou_threshold, normalize_means, normalize_stds): """Compute bounding-box regression targets for an image.""" if len(gt_rois) == 0: # Bail if the image has no ground-truth ROIs return np.zeros((proposals.shape[0], 6), dtype=np.float32) # Get IoU overlap between each ex ROI and gt ROI ex_gt_overlaps = bbox_overlaps( np.ascontiguousarray(proposals, dtype=np.float), np.ascontiguousarray(gt_rois, dtype=np.float)) # Indices of examples for which we try to make predictions ex_inds = np.where( ex_gt_overlaps >= iou_threshold)[0] # cfg.TRAIN.BBOX_THRESH # Find which gt ROI each ex ROI has max overlap with: # this will be the ex ROI's gt target gt_assignment_inds = ex_gt_overlaps.argmax(axis=1) gt_assignment_rois = gt_rois[gt_assignment_inds, :] regression_targets = bbox_transform(proposals[ex_inds], gt_assignment_rois[ex_inds]) # Optionally normalize targets by a precomputed mean and stdev if normalize_means is not None: regression_targets = (regression_targets - normalize_means) / normalize_stds targets = np.zeros((proposals.shape[0], 6), dtype=np.float32) targets[ex_inds, :4] = regression_targets targets[ex_inds, 4] = gt_rois[gt_assignment_inds[ex_inds], 4] targets[ex_inds, 5] = 1 # bbiw return targets
def compute_targets(proposals, gt_rois, iou_threshold, normalize_means, normalize_stds): """Compute bounding-box regression targets for an image.""" if len(gt_rois) == 0: # Bail if the image has no ground-truth ROIs return np.zeros((proposals.shape[0], 6), dtype=np.float32) # Get IoU overlap between each ex ROI and gt ROI ex_gt_overlaps = bbox_overlaps( np.ascontiguousarray(proposals, dtype=np.float), np.ascontiguousarray(gt_rois, dtype=np.float)) # Indices of examples for which we try to make predictions ex_inds = np.where(ex_gt_overlaps >= iou_threshold)[0] # cfg.TRAIN.BBOX_THRESH # Find which gt ROI each ex ROI has max overlap with: # this will be the ex ROI's gt target gt_assignment_inds = ex_gt_overlaps.argmax(axis=1) gt_assignment_rois = gt_rois[gt_assignment_inds, :] regression_targets = bbox_transform(proposals[ex_inds], gt_assignment_rois[ex_inds]) # Optionally normalize targets by a precomputed mean and stdev if normalize_means is not None: regression_targets = (regression_targets - normalize_means) / normalize_stds targets = np.zeros((proposals.shape[0], 6), dtype=np.float32) targets[ex_inds, :4] = regression_targets targets[ex_inds, 4] = gt_rois[gt_assignment_inds[ex_inds], 4] targets[ex_inds, 5] = 1 # bbiw return targets
def _sample_rois(all_rois, gt_boxes, fg_rois_per_image, rois_per_image, num_classes, deterministic=False): """Generate a random sample of RoIs comprising foreground and background examples. """ # overlaps: (rois x gt_boxes) overlaps = bbox_overlaps( np.ascontiguousarray(all_rois[:, 1:5], dtype=np.float), np.ascontiguousarray(gt_boxes[:, :4], dtype=np.float)) gt_assignment = overlaps.argmax(axis=1) max_overlaps = overlaps.max(axis=1) labels = gt_boxes[gt_assignment, 4] # Select foreground RoIs as those with >= FG_THRESH overlap fg_inds = np.where(max_overlaps >= cfg.TRAIN.FG_THRESH)[0] # Guard against the case when an image has fewer than fg_rois_per_image # foreground RoIs fg_rois_per_this_image = min(fg_rois_per_image, fg_inds.size) # Sample foreground regions without replacement if fg_inds.size > 0: if deterministic: fg_inds = fg_inds[:fg_rois_per_this_image] else: fg_inds = npr.choice(fg_inds, size=fg_rois_per_this_image, replace=False) # Select background RoIs as those within [BG_THRESH_LO, BG_THRESH_HI) bg_inds = np.where((max_overlaps < cfg.TRAIN.BG_THRESH_HI) & (max_overlaps >= cfg.TRAIN.BG_THRESH_LO))[0] # Compute number of background RoIs to take from this image (guarding # against there being fewer than desired) bg_rois_per_this_image = rois_per_image - fg_rois_per_this_image bg_rois_per_this_image = min(bg_rois_per_this_image, bg_inds.size) # Sample background regions without replacement if bg_inds.size > 0: if deterministic: bg_inds = bg_inds[:bg_rois_per_this_image] else: bg_inds = npr.choice(bg_inds, size=bg_rois_per_this_image, replace=False) # The indices that we're selecting (both fg and bg) keep_inds = np.append(fg_inds, bg_inds) # Select sampled values from various arrays: labels = labels[keep_inds] # Clamp labels for the background RoIs to 0 labels[fg_rois_per_this_image:] = 0 rois = all_rois[keep_inds] bbox_target_data = _compute_targets( rois[:, 1:5], gt_boxes[gt_assignment[keep_inds], :4], labels) bbox_targets, bbox_inside_weights = \ _get_bbox_regression_labels(bbox_target_data, num_classes) return labels, rois, bbox_targets, bbox_inside_weights
def _sample_rois(all_rois, gt_boxes, fg_rois_per_image, rois_per_image, num_classes, deterministic=False): """Generate a random sample of RoIs comprising foreground and background examples. """ # overlaps: (rois x gt_boxes) overlaps = bbox_overlaps( np.ascontiguousarray(all_rois[:, 1:5], dtype=np.float), np.ascontiguousarray(gt_boxes[:, :4], dtype=np.float)) gt_assignment = overlaps.argmax(axis=1) max_overlaps = overlaps.max(axis=1) labels = gt_boxes[gt_assignment, 4] # Select foreground RoIs as those with >= FG_THRESH overlap fg_inds = np.where(max_overlaps >= cfg["TRAIN"].FG_THRESH)[0] # Guard against the case when an image has fewer than fg_rois_per_image # foreground RoIs fg_rois_per_this_image = min(fg_rois_per_image, fg_inds.size) # Sample foreground regions without replacement if fg_inds.size > 0: if deterministic: fg_inds = fg_inds[:fg_rois_per_this_image] else: fg_inds = npr.choice(fg_inds, size=fg_rois_per_this_image, replace=False) # Select background RoIs as those within [BG_THRESH_LO, BG_THRESH_HI) bg_inds = np.where((max_overlaps < cfg["TRAIN"].BG_THRESH_HI) & (max_overlaps >= cfg["TRAIN"].BG_THRESH_LO))[0] # Compute number of background RoIs to take from this image (guarding # against there being fewer than desired) bg_rois_per_this_image = rois_per_image - fg_rois_per_this_image bg_rois_per_this_image = min(bg_rois_per_this_image, bg_inds.size) # Sample background regions without replacement if bg_inds.size > 0: if deterministic: bg_inds = bg_inds[:bg_rois_per_this_image] else: bg_inds = npr.choice(bg_inds, size=bg_rois_per_this_image, replace=False) # The indices that we're selecting (both fg and bg) keep_inds = np.append(fg_inds, bg_inds) # Select sampled values from various arrays: labels = labels[keep_inds] # Clamp labels for the background RoIs to 0 labels[fg_rois_per_this_image:] = 0 rois = all_rois[keep_inds] bbox_target_data = _compute_targets( rois[:, 1:5], gt_boxes[gt_assignment[keep_inds], :4], labels) bbox_targets, bbox_inside_weights = \ _get_bbox_regression_labels(bbox_target_data, num_classes) return labels, rois, bbox_targets, bbox_inside_weights
def forward(self, arguments, outputs, device=None, outputs_to_retain=None): # Algorithm: # # for each (H, W) location i # generate 9 anchor boxes centered on cell i # apply predicted bbox deltas at cell i to each of the 9 anchors # filter out-of-image anchors # measure GT overlap bottom = arguments # map of shape (..., H, W) height, width = bottom[0].shape[-2:] # GT boxes (x1, y1, x2, y2, label) gt_boxes = bottom[1][0,:] # im_info im_info = bottom[2] # remove zero padded ground truth boxes keep = np.where( ((gt_boxes[:,2] - gt_boxes[:,0]) > 0) & ((gt_boxes[:,3] - gt_boxes[:,1]) > 0) ) gt_boxes = gt_boxes[keep] if DEBUG: print ('') print ('im_size: ({}, {})'.format(im_info[0], im_info[1])) print ('scale: {}'.format(im_info[2])) print ('height, width: ({}, {})'.format(height, width)) print ('rpn: gt_boxes.shape', gt_boxes.shape) #print ('rpn: gt_boxes', gt_boxes) # 1. Generate proposals from bbox deltas and shifted anchors shift_x = np.arange(0, width) * self._feat_stride shift_y = np.arange(0, height) * self._feat_stride shift_x, shift_y = np.meshgrid(shift_x, shift_y) shifts = np.vstack((shift_x.ravel(), shift_y.ravel(), shift_x.ravel(), shift_y.ravel())).transpose() # add A anchors (1, A, 4) to # cell K shifts (K, 1, 4) to get # shift anchors (K, A, 4) # reshape to (K*A, 4) shifted anchors A = self._num_anchors K = shifts.shape[0] all_anchors = (self._anchors.reshape((1, A, 4)) + shifts.reshape((1, K, 4)).transpose((1, 0, 2))) all_anchors = all_anchors.reshape((K * A, 4)) total_anchors = int(K * A) # only keep anchors inside the image inds_inside = np.where( (all_anchors[:, 0] >= -self._allowed_border) & (all_anchors[:, 1] >= -self._allowed_border) & (all_anchors[:, 2] < im_info[1] + self._allowed_border) & # width (all_anchors[:, 3] < im_info[0] + self._allowed_border) # height )[0] if DEBUG: print ('total_anchors', total_anchors) print ('inds_inside', len(inds_inside)) # keep only inside anchors anchors = all_anchors[inds_inside, :] if DEBUG: print ('anchors.shape', anchors.shape) # label: 1 is positive, 0 is negative, -1 is dont care labels = np.empty((len(inds_inside), ), dtype=np.float32) labels.fill(-1) # overlaps between the anchors and the gt boxes # overlaps (ex, gt) overlaps = bbox_overlaps( np.ascontiguousarray(anchors, dtype=np.float), np.ascontiguousarray(gt_boxes, dtype=np.float)) argmax_overlaps = overlaps.argmax(axis=1) max_overlaps = overlaps[np.arange(len(inds_inside)), argmax_overlaps] gt_argmax_overlaps = overlaps.argmax(axis=0) gt_max_overlaps = overlaps[gt_argmax_overlaps, np.arange(overlaps.shape[1])] gt_argmax_overlaps = np.where(overlaps == gt_max_overlaps)[0] if not cfg["TRAIN"].RPN_CLOBBER_POSITIVES: # assign bg labels first so that positive labels can clobber them labels[max_overlaps < cfg["TRAIN"].RPN_NEGATIVE_OVERLAP] = 0 # fg label: for each gt, anchor with highest overlap labels[gt_argmax_overlaps] = 1 # fg label: above threshold IOU labels[max_overlaps >= cfg["TRAIN"].RPN_POSITIVE_OVERLAP] = 1 if cfg["TRAIN"].RPN_CLOBBER_POSITIVES: # assign bg labels last so that negative labels can clobber positives labels[max_overlaps < cfg["TRAIN"].RPN_NEGATIVE_OVERLAP] = 0 # subsample positive labels if we have too many num_fg = int(cfg["TRAIN"].RPN_FG_FRACTION * cfg["TRAIN"].RPN_BATCHSIZE) fg_inds = np.where(labels == 1)[0] if len(fg_inds) > num_fg: if self._determininistic_mode: disable_inds = fg_inds[:(len(fg_inds) - num_fg)] else: disable_inds = npr.choice(fg_inds, size=(len(fg_inds) - num_fg), replace=False) labels[disable_inds] = -1 # subsample negative labels if we have too many num_bg = cfg["TRAIN"].RPN_BATCHSIZE - np.sum(labels == 1) bg_inds = np.where(labels == 0)[0] if len(bg_inds) > num_bg: if self._determininistic_mode: disable_inds = bg_inds[:(len(bg_inds) - num_bg)] else: disable_inds = npr.choice(bg_inds, size=(len(bg_inds) - num_bg), replace=False) labels[disable_inds] = -1 bbox_targets = np.zeros((len(inds_inside), 4), dtype=np.float32) bbox_targets = _compute_targets(anchors, gt_boxes[argmax_overlaps, :]) bbox_inside_weights = np.zeros((len(inds_inside), 4), dtype=np.float32) bbox_inside_weights[labels == 1, :] = np.array((1.0, 1.0, 1.0, 1.0)) if DEBUG: self._sums += bbox_targets[labels == 1, :].sum(axis=0) self._squared_sums += (bbox_targets[labels == 1, :] ** 2).sum(axis=0) self._counts += np.sum(labels == 1) means = self._sums / self._counts stds = np.sqrt(self._squared_sums / self._counts - means ** 2) print ('means:') print (means) print ('stdevs:') print (stds) # map up to original set of anchors labels = _unmap(labels, total_anchors, inds_inside, fill=-1) bbox_targets = _unmap(bbox_targets, total_anchors, inds_inside, fill=0) bbox_inside_weights = _unmap(bbox_inside_weights, total_anchors, inds_inside, fill=0) if DEBUG: print ('rpn: max max_overlap', np.max(max_overlaps)) print ('rpn: num_positive', np.sum(labels == 1)) print ('rpn: num_negative', np.sum(labels == 0)) self._fg_sum += np.sum(labels == 1) self._bg_sum += np.sum(labels == 0) self._count += 1 print ('rpn: num_positive avg', self._fg_sum / self._count) print ('rpn: num_negative avg', self._bg_sum / self._count) # labels labels = labels.reshape((1, height, width, A)).transpose(0, 3, 1, 2) outputs[self.outputs[0]] = np.ascontiguousarray(labels) # bbox_targets bbox_targets = bbox_targets.reshape((1, height, width, A * 4)).transpose(0, 3, 1, 2) outputs[self.outputs[1]] = np.ascontiguousarray(bbox_targets) # bbox_inside_weights bbox_inside_weights = bbox_inside_weights \ .reshape((1, height, width, A * 4)).transpose(0, 3, 1, 2) assert bbox_inside_weights.shape[2] == height assert bbox_inside_weights.shape[3] == width outputs[self.outputs[2]] = np.ascontiguousarray(bbox_inside_weights) # No state needs to be passed to backward() so we just pass None return None
def forward(self, bottom, top): # Algorithm: # # for each (H, W) location i # generate 9 anchor boxes centered on cell i # apply predicted bbox deltas at cell i to each of the 9 anchors # filter out-of-image anchors # measure GT overlap assert bottom[0].data.shape[0] == 1, \ 'Only single item batches are supported' # map of shape (..., H, W) height, width = bottom[0].data.shape[-2:] # GT boxes (x1, y1, x2, y2, label) gt_boxes = bottom[1]#.data # im_info im_info = bottom[2]#.data[0, :] if DEBUG: print('') print('im_size: ({}, {})'.format(im_info[0], im_info[1])) print('scale: {}'.format(im_info[2])) print('height, width: ({}, {})'.format(height, width)) print('rpn: gt_boxes.shape', gt_boxes.shape) print('rpn: gt_boxes', gt_boxes) # 1. Generate proposals from bbox deltas and shifted anchors shift_x = np.arange(0, width) * self._feat_stride shift_y = np.arange(0, height) * self._feat_stride shift_x, shift_y = np.meshgrid(shift_x, shift_y) shifts = np.vstack((shift_x.ravel(), shift_y.ravel(), shift_x.ravel(), shift_y.ravel())).transpose() # add A anchors (1, A, 4) to # cell K shifts (K, 1, 4) to get # shift anchors (K, A, 4) # reshape to (K*A, 4) shifted anchors A = self._num_anchors K = shifts.shape[0] all_anchors = (self._anchors.reshape((1, A, 4)) + shifts.reshape((1, K, 4)).transpose((1, 0, 2))) all_anchors = all_anchors.reshape((K * A, 4)) total_anchors = int(K * A) # only keep anchors inside the image inds_inside = np.where( (all_anchors[:, 0] >= -self._allowed_border) & (all_anchors[:, 1] >= -self._allowed_border) & (all_anchors[:, 2] < im_info[1] + self._allowed_border) & # width (all_anchors[:, 3] < im_info[0] + self._allowed_border) # height )[0] if DEBUG: print('total_anchors', total_anchors) print('inds_inside', len(inds_inside)) # keep only inside anchors anchors = all_anchors[inds_inside, :] if DEBUG: print('anchors.shape', anchors.shape) # label: 1 is positive, 0 is negative, -1 is dont care labels = np.empty((len(inds_inside), ), dtype=np.float32) labels.fill(-1) # overlaps between the anchors and the gt boxes # overlaps (ex, gt) overlaps = bbox_overlaps( np.ascontiguousarray(anchors, dtype=np.float), np.ascontiguousarray(gt_boxes, dtype=np.float)) argmax_overlaps = overlaps.argmax(axis=1) max_overlaps = overlaps[np.arange(len(inds_inside)), argmax_overlaps] gt_argmax_overlaps = overlaps.argmax(axis=0) gt_max_overlaps = overlaps[gt_argmax_overlaps, np.arange(overlaps.shape[1])] gt_argmax_overlaps = np.where(overlaps == gt_max_overlaps)[0] if not cfg.TRAIN.RPN_CLOBBER_POSITIVES: # assign bg labels first so that positive labels can clobber them labels[max_overlaps < cfg.TRAIN.RPN_NEGATIVE_OVERLAP] = 0 # fg label: for each gt, anchor with highest overlap labels[gt_argmax_overlaps] = 1 # fg label: above threshold IOU labels[max_overlaps >= cfg.TRAIN.RPN_POSITIVE_OVERLAP] = 1 if cfg.TRAIN.RPN_CLOBBER_POSITIVES: # assign bg labels last so that negative labels can clobber positives labels[max_overlaps < cfg.TRAIN.RPN_NEGATIVE_OVERLAP] = 0 # subsample positive labels if we have too many num_fg = int(cfg.TRAIN.RPN_FG_FRACTION * cfg.TRAIN.RPN_BATCHSIZE) fg_inds = np.where(labels == 1)[0] if len(fg_inds) > num_fg: if self._determininistic_mode: disable_inds = fg_inds[:(len(fg_inds) - num_fg)] else: disable_inds = npr.choice(fg_inds, size=(len(fg_inds) - num_fg), replace=False) labels[disable_inds] = -1 # subsample negative labels if we have too many num_bg = cfg.TRAIN.RPN_BATCHSIZE - np.sum(labels == 1) bg_inds = np.where(labels == 0)[0] if len(bg_inds) > num_bg: if self._determininistic_mode: disable_inds = bg_inds[:(len(bg_inds) - num_bg)] else: disable_inds = npr.choice(bg_inds, size=(len(bg_inds) - num_bg), replace=False) labels[disable_inds] = -1 #print "was %s inds, disabling %s, now %s inds" % ( #len(bg_inds), len(disable_inds), np.sum(labels == 0)) bbox_targets = np.zeros((len(inds_inside), 4), dtype=np.float32) bbox_targets = _compute_targets(anchors, gt_boxes[argmax_overlaps, :]) bbox_inside_weights = np.zeros((len(inds_inside), 4), dtype=np.float32) bbox_inside_weights[labels == 1, :] = np.array(cfg.TRAIN.RPN_BBOX_INSIDE_WEIGHTS) bbox_outside_weights = np.zeros((len(inds_inside), 4), dtype=np.float32) if cfg.TRAIN.RPN_POSITIVE_WEIGHT < 0: # uniform weighting of examples (given non-uniform sampling) num_examples = np.sum(labels >= 0) positive_weights = np.ones((1, 4)) * 1.0 / num_examples negative_weights = np.ones((1, 4)) * 1.0 / num_examples else: assert ((cfg.TRAIN.RPN_POSITIVE_WEIGHT > 0) & (cfg.TRAIN.RPN_POSITIVE_WEIGHT < 1)) positive_weights = (cfg.TRAIN.RPN_POSITIVE_WEIGHT / np.sum(labels == 1)) negative_weights = ((1.0 - cfg.TRAIN.RPN_POSITIVE_WEIGHT) / np.sum(labels == 0)) bbox_outside_weights[labels == 1, :] = positive_weights bbox_outside_weights[labels == 0, :] = negative_weights if DEBUG: self._sums += bbox_targets[labels == 1, :].sum(axis=0) self._squared_sums += (bbox_targets[labels == 1, :] ** 2).sum(axis=0) self._counts += np.sum(labels == 1) means = self._sums / self._counts stds = np.sqrt(self._squared_sums / self._counts - means ** 2) print('means:') print(means) print('stdevs:') print(stds) # map up to original set of anchors labels = _unmap(labels, total_anchors, inds_inside, fill=-1) bbox_targets = _unmap(bbox_targets, total_anchors, inds_inside, fill=0) bbox_inside_weights = _unmap(bbox_inside_weights, total_anchors, inds_inside, fill=0) bbox_outside_weights = _unmap(bbox_outside_weights, total_anchors, inds_inside, fill=0) if DEBUG: print('rpn: max max_overlap', np.max(max_overlaps)) print('rpn: num_positive', np.sum(labels == 1)) print('rpn: num_negative', np.sum(labels == 0)) self._fg_sum += np.sum(labels == 1) self._bg_sum += np.sum(labels == 0) self._count += 1 print('rpn: num_positive avg', self._fg_sum / self._count) print('rpn: num_negative avg', self._bg_sum / self._count) # labels labels = labels.reshape((1, height, width, A)).transpose(0, 3, 1, 2) #labels = labels.reshape((1, 1, A * height, width)) #top[0].reshape(*labels.shape) #top[0].data[...] = labels # bbox_targets bbox_targets = bbox_targets \ .reshape((1, height, width, A * 4)).transpose(0, 3, 1, 2) #top[1].reshape(*bbox_targets.shape) #top[1].data[...] = bbox_targets # bbox_inside_weights bbox_inside_weights = bbox_inside_weights \ .reshape((1, height, width, A * 4)).transpose(0, 3, 1, 2) #assert bbox_inside_weights.shape[2] == height #assert bbox_inside_weights.shape[3] == width #top[2].reshape(*bbox_inside_weights.shape) #top[2].data[...] = bbox_inside_weights # bbox_outside_weights #bbox_outside_weights = bbox_outside_weights \ # .reshape((1, height, width, A * 4)).transpose(0, 3, 1, 2) #assert bbox_outside_weights.shape[2] == height #assert bbox_outside_weights.shape[3] == width #top[3].reshape(*bbox_outside_weights.shape) #top[3].data[...] = bbox_outside_weights return labels, bbox_targets, bbox_inside_weights
def forward(self, bottom, top): # Algorithm: # # for each (H, W) location i # generate 9 anchor boxes centered on cell i # apply predicted bbox deltas at cell i to each of the 9 anchors # filter out-of-image anchors # measure GT overlap assert bottom[0].data.shape[0] == 1, \ 'Only single item batches are supported' # map of shape (..., H, W) height, width = bottom[0].data.shape[-2:] # GT boxes (x1, y1, x2, y2, label) gt_boxes = bottom[1]#.data # im_info im_info = bottom[2]#.data[0, :] if DEBUG: print('') print('im_size: ({}, {})'.format(im_info[0], im_info[1])) print('scale: {}'.format(im_info[2])) print('height, width: ({}, {})'.format(height, width)) print('rpn: gt_boxes.shape', gt_boxes.shape) print('rpn: gt_boxes', gt_boxes) # 1. Generate proposals from bbox deltas and shifted anchors shift_x = np.arange(0, width) * self._feat_stride shift_y = np.arange(0, height) * self._feat_stride shift_x, shift_y = np.meshgrid(shift_x, shift_y) shifts = np.vstack((shift_x.ravel(), shift_y.ravel(), shift_x.ravel(), shift_y.ravel())).transpose() # add A anchors (1, A, 4) to # cell K shifts (K, 1, 4) to get # shift anchors (K, A, 4) # reshape to (K*A, 4) shifted anchors A = self._num_anchors K = shifts.shape[0] all_anchors = (self._anchors.reshape((1, A, 4)) + shifts.reshape((1, K, 4)).transpose((1, 0, 2))) all_anchors = all_anchors.reshape((K * A, 4)) total_anchors = int(K * A) # only keep anchors inside the image inds_inside = np.where( (all_anchors[:, 0] >= -self._allowed_border) & (all_anchors[:, 1] >= -self._allowed_border) & (all_anchors[:, 2] < im_info[1] + self._allowed_border) & # width (all_anchors[:, 3] < im_info[0] + self._allowed_border) # height )[0] if DEBUG: print('total_anchors', total_anchors) print('inds_inside', len(inds_inside)) # keep only inside anchors anchors = all_anchors[inds_inside, :] if DEBUG: print('anchors.shape', anchors.shape) # label: 1 is positive, 0 is negative, -1 is dont care labels = np.empty((len(inds_inside), ), dtype=np.float32) labels.fill(-1) # overlaps between the anchors and the gt boxes # overlaps (ex, gt) overlaps = bbox_overlaps( np.ascontiguousarray(anchors, dtype=np.float), np.ascontiguousarray(gt_boxes, dtype=np.float)) argmax_overlaps = overlaps.argmax(axis=1) max_overlaps = overlaps[np.arange(len(inds_inside)), argmax_overlaps] gt_argmax_overlaps = overlaps.argmax(axis=0) gt_max_overlaps = overlaps[gt_argmax_overlaps, np.arange(overlaps.shape[1])] gt_argmax_overlaps = np.where(overlaps == gt_max_overlaps)[0] if not cfg.TRAIN.RPN_CLOBBER_POSITIVES: # assign bg labels first so that positive labels can clobber them labels[max_overlaps < cfg.TRAIN.RPN_NEGATIVE_OVERLAP] = 0 # fg label: for each gt, anchor with highest overlap labels[gt_argmax_overlaps] = 1 # fg label: above threshold IOU labels[max_overlaps >= cfg.TRAIN.RPN_POSITIVE_OVERLAP] = 1 if cfg.TRAIN.RPN_CLOBBER_POSITIVES: # assign bg labels last so that negative labels can clobber positives labels[max_overlaps < cfg.TRAIN.RPN_NEGATIVE_OVERLAP] = 0 # subsample positive labels if we have too many num_fg = int(cfg.TRAIN.RPN_FG_FRACTION * cfg.TRAIN.RPN_BATCHSIZE) fg_inds = np.where(labels == 1)[0] if len(fg_inds) > num_fg: if self._determininistic_mode: disable_inds = fg_inds[:(len(fg_inds) - num_fg)] else: disable_inds = npr.choice(fg_inds, size=(len(fg_inds) - num_fg), replace=False) labels[disable_inds] = -1 # subsample negative labels if we have too many num_bg = cfg.TRAIN.RPN_BATCHSIZE - np.sum(labels == 1) bg_inds = np.where(labels == 0)[0] if len(bg_inds) > num_bg: if self._determininistic_mode: disable_inds = bg_inds[:(len(bg_inds) - num_bg)] else: disable_inds = npr.choice(bg_inds, size=(len(bg_inds) - num_bg), replace=False) labels[disable_inds] = -1 #print "was %s inds, disabling %s, now %s inds" % ( #len(bg_inds), len(disable_inds), np.sum(labels == 0)) bbox_targets = np.zeros((len(inds_inside), 4), dtype=np.float32) bbox_targets = _compute_targets(anchors, gt_boxes[argmax_overlaps, :]) bbox_inside_weights = np.zeros((len(inds_inside), 4), dtype=np.float32) bbox_inside_weights[labels == 1, :] = np.array(cfg.TRAIN.RPN_BBOX_INSIDE_WEIGHTS) bbox_outside_weights = np.zeros((len(inds_inside), 4), dtype=np.float32) if cfg.TRAIN.RPN_POSITIVE_WEIGHT < 0: # uniform weighting of examples (given non-uniform sampling) num_examples = np.sum(labels >= 0) positive_weights = np.ones((1, 4)) * 1.0 / num_examples negative_weights = np.ones((1, 4)) * 1.0 / num_examples else: assert ((cfg.TRAIN.RPN_POSITIVE_WEIGHT > 0) & (cfg.TRAIN.RPN_POSITIVE_WEIGHT < 1)) positive_weights = (cfg.TRAIN.RPN_POSITIVE_WEIGHT / np.sum(labels == 1)) negative_weights = ((1.0 - cfg.TRAIN.RPN_POSITIVE_WEIGHT) / np.sum(labels == 0)) bbox_outside_weights[labels == 1, :] = positive_weights bbox_outside_weights[labels == 0, :] = negative_weights if DEBUG: self._sums += bbox_targets[labels == 1, :].sum(axis=0) self._squared_sums += (bbox_targets[labels == 1, :] ** 2).sum(axis=0) self._counts += np.sum(labels == 1) means = self._sums / self._counts stds = np.sqrt(self._squared_sums / self._counts - means ** 2) print('means:') print(means) print('stdevs:') print(stds) # map up to original set of anchors labels = _unmap(labels, total_anchors, inds_inside, fill=-1) bbox_targets = _unmap(bbox_targets, total_anchors, inds_inside, fill=0) bbox_inside_weights = _unmap(bbox_inside_weights, total_anchors, inds_inside, fill=0) bbox_outside_weights = _unmap(bbox_outside_weights, total_anchors, inds_inside, fill=0) if DEBUG: print('rpn: max max_overlap', np.max(max_overlaps)) print('rpn: num_positive', np.sum(labels == 1)) print('rpn: num_negative', np.sum(labels == 0)) self._fg_sum += np.sum(labels == 1) self._bg_sum += np.sum(labels == 0) self._count += 1 print('rpn: num_positive avg', self._fg_sum / self._count) print('rpn: num_negative avg', self._bg_sum / self._count) # labels labels = labels.reshape((1, height, width, A)).transpose(0, 3, 1, 2) #labels = labels.reshape((1, 1, A * height, width)) #top[0].reshape(*labels.shape) #top[0].data[...] = labels # bbox_targets bbox_targets = bbox_targets \ .reshape((1, height, width, A * 4)).transpose(0, 3, 1, 2) #top[1].reshape(*bbox_targets.shape) #top[1].data[...] = bbox_targets # bbox_inside_weights bbox_inside_weights = bbox_inside_weights \ .reshape((1, height, width, A * 4)).transpose(0, 3, 1, 2) #assert bbox_inside_weights.shape[2] == height #assert bbox_inside_weights.shape[3] == width #top[2].reshape(*bbox_inside_weights.shape) #top[2].data[...] = bbox_inside_weights # bbox_outside_weights #bbox_outside_weights = bbox_outside_weights \ # .reshape((1, height, width, A * 4)).transpose(0, 3, 1, 2) #assert bbox_outside_weights.shape[2] == height #assert bbox_outside_weights.shape[3] == width #top[3].reshape(*bbox_outside_weights.shape) #top[3].data[...] = bbox_outside_weights return labels, bbox_targets, bbox_inside_weights
def forward(self, arguments, outputs, device=None, outputs_to_retain=None): # Algorithm: # # for each (H, W) location i # generate 9 anchor boxes centered on cell i # apply predicted bbox deltas at cell i to each of the 9 anchors # filter out-of-image anchors # measure GT overlap bottom = arguments # map of shape (..., H, W) height, width = bottom[0].shape[-2:] # GT boxes (x1, y1, x2, y2, label) gt_boxes = bottom[1][0, :] # im_info im_info = bottom[2][0] # remove zero padded ground truth boxes keep = np.where(((gt_boxes[:, 2] - gt_boxes[:, 0]) > 0) & ((gt_boxes[:, 3] - gt_boxes[:, 1]) > 0)) gt_boxes = gt_boxes[keep] if DEBUG: print('') # im_info = (pad_width, pad_height, scaled_image_width, scaled_image_height, orig_img_width, orig_img_height) # e.g.(1000, 1000, 1000, 600, 500, 300) for an original image of 600x300 that is scaled and padded to 1000x1000 print('im_size: ({}, {})'.format(im_info[0], im_info[1])) print('scaled im_size: ({}, {})'.format(im_info[2], im_info[3])) print('original im_size: ({}, {})'.format(im_info[4], im_info[5])) print('height, width: ({}, {})'.format(height, width)) print('rpn: gt_boxes.shape', gt_boxes.shape) # print ('rpn: gt_boxes', gt_boxes) # 1. Generate proposals from bbox deltas and shifted anchors shift_x = np.arange(0, width) * self._feat_stride shift_y = np.arange(0, height) * self._feat_stride shift_x, shift_y = np.meshgrid(shift_x, shift_y) shifts = np.vstack((shift_x.ravel(), shift_y.ravel(), shift_x.ravel(), shift_y.ravel())).transpose() # add A anchors (1, A, 4) to # cell K shifts (K, 1, 4) to get # shift anchors (K, A, 4) # reshape to (K*A, 4) shifted anchors A = self._num_anchors K = shifts.shape[0] all_anchors = (self._anchors.reshape((1, A, 4)) + shifts.reshape( (1, K, 4)).transpose((1, 0, 2))) all_anchors = all_anchors.reshape((K * A, 4)) total_anchors = int(K * A) # only keep anchors inside the image padded_wh = im_info[0:2] scaled_wh = im_info[2:4] xy_offset = (padded_wh - scaled_wh) / 2 xy_min = xy_offset xy_max = xy_offset + scaled_wh inds_inside = np.where( (all_anchors[:, 0] >= xy_min[0] - self._allowed_border) & (all_anchors[:, 1] >= xy_min[1] - self._allowed_border) & (all_anchors[:, 2] < xy_max[0] + self._allowed_border) & # width (all_anchors[:, 3] < xy_max[1] + self._allowed_border) # height )[0] if DEBUG: print('total_anchors', total_anchors) print('inds_inside', len(inds_inside)) # keep only inside anchors anchors = all_anchors[inds_inside, :] if DEBUG: print('anchors.shape', anchors.shape) print('gt_boxes.shape', gt_boxes.shape) # label: 1 is positive, 0 is negative, -1 is dont care labels = np.empty((len(inds_inside), ), dtype=np.float32) labels.fill(-1) # overlaps between the anchors and the gt boxes # overlaps (ex, gt) overlaps = bbox_overlaps( np.ascontiguousarray(anchors, dtype=np.float), np.ascontiguousarray(gt_boxes, dtype=np.float)) argmax_overlaps = overlaps.argmax(axis=1) max_overlaps = overlaps[np.arange(len(inds_inside)), argmax_overlaps] gt_argmax_overlaps = overlaps.argmax(axis=0) gt_max_overlaps = overlaps[gt_argmax_overlaps, np.arange(overlaps.shape[1])] gt_argmax_overlaps = np.where(overlaps == gt_max_overlaps)[0] if not self._clobber_positives: # assign bg labels first so that positive labels can clobber them labels[max_overlaps < self._negative_overlap] = 0 # fg label: for each gt, anchor with highest overlap labels[gt_argmax_overlaps] = 1 # fg label: above threshold IOU labels[max_overlaps >= self._positive_overlap] = 1 if self._clobber_positives: # assign bg labels last so that negative labels can clobber positives labels[max_overlaps < self._negative_overlap] = 0 # subsample positive labels if we have too many num_fg = int(self._rpn_fg_fraction * self._rpn_batch_size) fg_inds = np.where(labels == 1)[0] if len(fg_inds) > num_fg: if self._determininistic_mode: disable_inds = fg_inds[:(len(fg_inds) - num_fg)] else: disable_inds = npr.choice(fg_inds, size=(len(fg_inds) - num_fg), replace=False) labels[disable_inds] = -1 # subsample negative labels if we have too many num_bg = self._rpn_batch_size - np.sum(labels == 1) bg_inds = np.where(labels == 0)[0] if len(bg_inds) > num_bg: if self._determininistic_mode: disable_inds = bg_inds[:(len(bg_inds) - num_bg)] else: disable_inds = npr.choice(bg_inds, size=(len(bg_inds) - num_bg), replace=False) labels[disable_inds] = -1 bbox_targets = np.zeros((len(inds_inside), 4), dtype=np.float32) bbox_targets = _compute_targets(anchors, gt_boxes[argmax_overlaps, :]) bbox_inside_weights = np.zeros((len(inds_inside), 4), dtype=np.float32) bbox_inside_weights[labels == 1, :] = np.array((1.0, 1.0, 1.0, 1.0)) if DEBUG: self._sums += bbox_targets[labels == 1, :].sum(axis=0) self._squared_sums += (bbox_targets[labels == 1, :]**2).sum(axis=0) self._counts += np.sum(labels == 1) means = self._sums / self._counts stds = np.sqrt(self._squared_sums / self._counts - means**2) print('means:') print(means) print('stdevs:') print(stds) # map up to original set of anchors labels = _unmap(labels, total_anchors, inds_inside, fill=-1) bbox_targets = _unmap(bbox_targets, total_anchors, inds_inside, fill=0) bbox_inside_weights = _unmap(bbox_inside_weights, total_anchors, inds_inside, fill=0) if DEBUG: print('rpn: max max_overlap', np.max(max_overlaps)) print('rpn: num_positive', np.sum(labels == 1)) print('rpn: num_negative', np.sum(labels == 0)) self._fg_sum += np.sum(labels == 1) self._bg_sum += np.sum(labels == 0) self._count += 1 print('rpn: num_positive avg', self._fg_sum / self._count) print('rpn: num_negative avg', self._bg_sum / self._count) # labels labels = labels.reshape((1, height, width, A)).transpose(0, 3, 1, 2) outputs[self.outputs[0]] = np.ascontiguousarray(labels) # bbox_targets bbox_targets = bbox_targets.reshape( (1, height, width, A * 4)).transpose(0, 3, 1, 2) outputs[self.outputs[1]] = np.ascontiguousarray(bbox_targets) # bbox_inside_weights bbox_inside_weights = bbox_inside_weights \ .reshape((1, height, width, A * 4)).transpose(0, 3, 1, 2) assert bbox_inside_weights.shape[2] == height assert bbox_inside_weights.shape[3] == width outputs[self.outputs[2]] = np.ascontiguousarray(bbox_inside_weights) # No state needs to be passed to backward() so we just pass None return None