mmdet/models/dense_heads/fovea_head.py

# Copyright (c) OpenMMLab. All rights reserved.
import warnings

import torch
import torch.nn as nn
from mmcv.cnn import ConvModule
from mmcv.ops import DeformConv2d
from mmcv.runner import BaseModule

from mmdet.core import multi_apply
from mmdet.core.utils import filter_scores_and_topk
from ..builder import HEADS
from .anchor_free_head import AnchorFreeHead

INF = 1e8


class FeatureAlign(BaseModule):

    def __init__(self,
                 in_channels,
                 out_channels,
                 kernel_size=3,
                 deform_groups=4,
                 init_cfg=dict(
                     type='Normal',
                     layer='Conv2d',
                     std=0.1,
                     override=dict(
                         type='Normal', name='conv_adaption', std=0.01))):
        super(FeatureAlign, self).__init__(init_cfg)
        offset_channels = kernel_size * kernel_size * 2
        self.conv_offset = nn.Conv2d(
            4, deform_groups * offset_channels, 1, bias=False)
        self.conv_adaption = DeformConv2d(
            in_channels,
            out_channels,
            kernel_size=kernel_size,
            padding=(kernel_size - 1) // 2,
            deform_groups=deform_groups)
        self.relu = nn.ReLU(inplace=True)

    def forward(self, x, shape):
        offset = self.conv_offset(shape)
        x = self.relu(self.conv_adaption(x, offset))
        return x


@HEADS.register_module()
class FoveaHead(AnchorFreeHead):
    """FoveaBox: Beyond Anchor-based Object Detector
    https://arxiv.org/abs/1904.03797
    """

    def __init__(self,
                 num_classes,
                 in_channels,
                 base_edge_list=(16, 32, 64, 128, 256),
                 scale_ranges=((8, 32), (16, 64), (32, 128), (64, 256), (128,
                                                                         512)),
                 sigma=0.4,
                 with_deform=False,
                 deform_groups=4,
                 init_cfg=dict(
                     type='Normal',
                     layer='Conv2d',
                     std=0.01,
                     override=dict(
                         type='Normal',
                         name='conv_cls',
                         std=0.01,
                         bias_prob=0.01)),
                 **kwargs):
        self.base_edge_list = base_edge_list
        self.scale_ranges = scale_ranges
        self.sigma = sigma
        self.with_deform = with_deform
        self.deform_groups = deform_groups
        super().__init__(num_classes, in_channels, init_cfg=init_cfg, **kwargs)

    def _init_layers(self):
        # box branch
        super()._init_reg_convs()
        self.conv_reg = nn.Conv2d(self.feat_channels, 4, 3, padding=1)

        # cls branch
        if not self.with_deform:
            super()._init_cls_convs()
            self.conv_cls = nn.Conv2d(
                self.feat_channels, self.cls_out_channels, 3, padding=1)
        else:
            self.cls_convs = nn.ModuleList()
            self.cls_convs.append(
                ConvModule(
                    self.feat_channels, (self.feat_channels * 4),
                    3,
                    stride=1,
                    padding=1,
                    conv_cfg=self.conv_cfg,
                    norm_cfg=self.norm_cfg,
                    bias=self.norm_cfg is None))
            self.cls_convs.append(
                ConvModule((self.feat_channels * 4), (self.feat_channels * 4),
                           1,
                           stride=1,
                           padding=0,
                           conv_cfg=self.conv_cfg,
                           norm_cfg=self.norm_cfg,
                           bias=self.norm_cfg is None))
            self.feature_adaption = FeatureAlign(
                self.feat_channels,
                self.feat_channels,
                kernel_size=3,
                deform_groups=self.deform_groups)
            self.conv_cls = nn.Conv2d(
                int(self.feat_channels * 4),
                self.cls_out_channels,
                3,
                padding=1)

    def forward_single(self, x):
        cls_feat = x
        reg_feat = x
        for reg_layer in self.reg_convs:
            reg_feat = reg_layer(reg_feat)
        bbox_pred = self.conv_reg(reg_feat)
        if self.with_deform:
            cls_feat = self.feature_adaption(cls_feat, bbox_pred.exp())
        for cls_layer in self.cls_convs:
            cls_feat = cls_layer(cls_feat)
        cls_score = self.conv_cls(cls_feat)
        return cls_score, bbox_pred

    def loss(self,
             cls_scores,
             bbox_preds,
             gt_bbox_list,
             gt_label_list,
             img_metas,
             gt_bboxes_ignore=None):
        assert len(cls_scores) == len(bbox_preds)

        featmap_sizes = [featmap.size()[-2:] for featmap in cls_scores]
        points = self.prior_generator.grid_priors(
            featmap_sizes,
            dtype=bbox_preds[0].dtype,
            device=bbox_preds[0].device)
        num_imgs = cls_scores[0].size(0)
        flatten_cls_scores = [
            cls_score.permute(0, 2, 3, 1).reshape(-1, self.cls_out_channels)
            for cls_score in cls_scores
        ]
        flatten_bbox_preds = [
            bbox_pred.permute(0, 2, 3, 1).reshape(-1, 4)
            for bbox_pred in bbox_preds
        ]
        flatten_cls_scores = torch.cat(flatten_cls_scores)
        flatten_bbox_preds = torch.cat(flatten_bbox_preds)
        flatten_labels, flatten_bbox_targets = self.get_targets(
            gt_bbox_list, gt_label_list, featmap_sizes, points)

        # FG cat_id: [0, num_classes -1], BG cat_id: num_classes
        pos_inds = ((flatten_labels >= 0)
                    & (flatten_labels < self.num_classes)).nonzero().view(-1)
        num_pos = len(pos_inds)

        loss_cls = self.loss_cls(
            flatten_cls_scores, flatten_labels, avg_factor=num_pos + num_imgs)
        if num_pos > 0:
            pos_bbox_preds = flatten_bbox_preds[pos_inds]
            pos_bbox_targets = flatten_bbox_targets[pos_inds]
            pos_weights = pos_bbox_targets.new_zeros(
                pos_bbox_targets.size()) + 1.0
            loss_bbox = self.loss_bbox(
                pos_bbox_preds,
                pos_bbox_targets,
                pos_weights,
                avg_factor=num_pos)
        else:
            loss_bbox = torch.tensor(
                0,
                dtype=flatten_bbox_preds.dtype,
                device=flatten_bbox_preds.device)
        return dict(loss_cls=loss_cls, loss_bbox=loss_bbox)

    def get_targets(self, gt_bbox_list, gt_label_list, featmap_sizes, points):
        label_list, bbox_target_list = multi_apply(
            self._get_target_single,
            gt_bbox_list,
            gt_label_list,
            featmap_size_list=featmap_sizes,
            point_list=points)
        flatten_labels = [
            torch.cat([
                labels_level_img.flatten() for labels_level_img in labels_level
            ]) for labels_level in zip(*label_list)
        ]
        flatten_bbox_targets = [
            torch.cat([
                bbox_targets_level_img.reshape(-1, 4)
                for bbox_targets_level_img in bbox_targets_level
            ]) for bbox_targets_level in zip(*bbox_target_list)
        ]
        flatten_labels = torch.cat(flatten_labels)
        flatten_bbox_targets = torch.cat(flatten_bbox_targets)
        return flatten_labels, flatten_bbox_targets

    def _get_target_single(self,
                           gt_bboxes_raw,
                           gt_labels_raw,
                           featmap_size_list=None,
                           point_list=None):

        gt_areas = torch.sqrt((gt_bboxes_raw[:, 2] - gt_bboxes_raw[:, 0]) *
                              (gt_bboxes_raw[:, 3] - gt_bboxes_raw[:, 1]))
        label_list = []
        bbox_target_list = []
        # for each pyramid, find the cls and box target
        for base_len, (lower_bound, upper_bound), stride, featmap_size, \
            points in zip(self.base_edge_list, self.scale_ranges,
                          self.strides, featmap_size_list, point_list):
            # FG cat_id: [0, num_classes -1], BG cat_id: num_classes
            points = points.view(*featmap_size, 2)
            x, y = points[..., 0], points[..., 1]
            labels = gt_labels_raw.new_zeros(featmap_size) + self.num_classes
            bbox_targets = gt_bboxes_raw.new(featmap_size[0], featmap_size[1],
                                             4) + 1
            # scale assignment
            hit_indices = ((gt_areas >= lower_bound) &
                           (gt_areas <= upper_bound)).nonzero().flatten()
            if len(hit_indices) == 0:
                label_list.append(labels)
                bbox_target_list.append(torch.log(bbox_targets))
                continue
            _, hit_index_order = torch.sort(-gt_areas[hit_indices])
            hit_indices = hit_indices[hit_index_order]
            gt_bboxes = gt_bboxes_raw[hit_indices, :] / stride
            gt_labels = gt_labels_raw[hit_indices]
            half_w = 0.5 * (gt_bboxes[:, 2] - gt_bboxes[:, 0])
            half_h = 0.5 * (gt_bboxes[:, 3] - gt_bboxes[:, 1])
            # valid fovea area: left, right, top, down
            pos_left = torch.ceil(
                gt_bboxes[:, 0] + (1 - self.sigma) * half_w - 0.5).long(). \
                clamp(0, featmap_size[1] - 1)
            pos_right = torch.floor(
                gt_bboxes[:, 0] + (1 + self.sigma) * half_w - 0.5).long(). \
                clamp(0, featmap_size[1] - 1)
            pos_top = torch.ceil(
                gt_bboxes[:, 1] + (1 - self.sigma) * half_h - 0.5).long(). \
                clamp(0, featmap_size[0] - 1)
            pos_down = torch.floor(
                gt_bboxes[:, 1] + (1 + self.sigma) * half_h - 0.5).long(). \
                clamp(0, featmap_size[0] - 1)
            for px1, py1, px2, py2, label, (gt_x1, gt_y1, gt_x2, gt_y2) in \
                    zip(pos_left, pos_top, pos_right, pos_down, gt_labels,
                        gt_bboxes_raw[hit_indices, :]):
                labels[py1:py2 + 1, px1:px2 + 1] = label
                bbox_targets[py1:py2 + 1, px1:px2 + 1, 0] = \
                    (x[py1:py2 + 1, px1:px2 + 1] - gt_x1) / base_len
                bbox_targets[py1:py2 + 1, px1:px2 + 1, 1] = \
                    (y[py1:py2 + 1, px1:px2 + 1] - gt_y1) / base_len
                bbox_targets[py1:py2 + 1, px1:px2 + 1, 2] = \
                    (gt_x2 - x[py1:py2 + 1, px1:px2 + 1]) / base_len
                bbox_targets[py1:py2 + 1, px1:px2 + 1, 3] = \
                    (gt_y2 - y[py1:py2 + 1, px1:px2 + 1]) / base_len
            bbox_targets = bbox_targets.clamp(min=1. / 16, max=16.)
            label_list.append(labels)
            bbox_target_list.append(torch.log(bbox_targets))
        return label_list, bbox_target_list

    # Same as base_dense_head/_get_bboxes_single except self._bbox_decode
    def _get_bboxes_single(self,
                           cls_score_list,
                           bbox_pred_list,
                           score_factor_list,
                           mlvl_priors,
                           img_meta,
                           cfg,
                           rescale=False,
                           with_nms=True,
                           **kwargs):
        """Transform outputs of a single image into bbox predictions.

        Args:
            cls_score_list (list[Tensor]): Box scores from all scale
                levels of a single image, each item has shape
                (num_priors * num_classes, H, W).
            bbox_pred_list (list[Tensor]): Box energies / deltas from
                all scale levels of a single image, each item has shape
                (num_priors * 4, H, W).
            score_factor_list (list[Tensor]): Score factor from all scale
                levels of a single image. Fovea head does not need this value.
            mlvl_priors (list[Tensor]): Each element in the list is
                the priors of a single level in feature pyramid, has shape
                (num_priors, 2).
            img_meta (dict): Image meta info.
            cfg (mmcv.Config): Test / postprocessing configuration,
                if None, test_cfg would be used.
            rescale (bool): If True, return boxes in original image space.
                Default: False.
            with_nms (bool): If True, do nms before return boxes.
                Default: True.

        Returns:
            tuple[Tensor]: Results of detected bboxes and labels. If with_nms
                is False and mlvl_score_factor is None, return mlvl_bboxes and
                mlvl_scores, else return mlvl_bboxes, mlvl_scores and
                mlvl_score_factor. Usually with_nms is False is used for aug
                test. If with_nms is True, then return the following format

                - det_bboxes (Tensor): Predicted bboxes with shape \
                    [num_bboxes, 5], where the first 4 columns are bounding \
                    box positions (tl_x, tl_y, br_x, br_y) and the 5-th \
                    column are scores between 0 and 1.
                - det_labels (Tensor): Predicted labels of the corresponding \
                    box with shape [num_bboxes].
        """
        cfg = self.test_cfg if cfg is None else cfg
        assert len(cls_score_list) == len(bbox_pred_list)
        img_shape = img_meta['img_shape']
        nms_pre = cfg.get('nms_pre', -1)

        mlvl_bboxes = []
        mlvl_scores = []
        mlvl_labels = []
        for level_idx, (cls_score, bbox_pred, stride, base_len, priors) in \
                enumerate(zip(cls_score_list, bbox_pred_list, self.strides,
                              self.base_edge_list, mlvl_priors)):
            assert cls_score.size()[-2:] == bbox_pred.size()[-2:]
            bbox_pred = bbox_pred.permute(1, 2, 0).reshape(-1, 4)

            scores = cls_score.permute(1, 2, 0).reshape(
                -1, self.cls_out_channels).sigmoid()

            # After https://github.com/open-mmlab/mmdetection/pull/6268/,
            # this operation keeps fewer bboxes under the same `nms_pre`.
            # There is no difference in performance for most models. If you
            # find a slight drop in performance, you can set a larger
            # `nms_pre` than before.
            results = filter_scores_and_topk(
                scores, cfg.score_thr, nms_pre,
                dict(bbox_pred=bbox_pred, priors=priors))
            scores, labels, _, filtered_results = results

            bbox_pred = filtered_results['bbox_pred']
            priors = filtered_results['priors']

            bboxes = self._bbox_decode(priors, bbox_pred, base_len, img_shape)

            mlvl_bboxes.append(bboxes)
            mlvl_scores.append(scores)
            mlvl_labels.append(labels)

        return self._bbox_post_process(mlvl_scores, mlvl_labels, mlvl_bboxes,
                                       img_meta['scale_factor'], cfg, rescale,
                                       with_nms)

    def _bbox_decode(self, priors, bbox_pred, base_len, max_shape):
        bbox_pred = bbox_pred.exp()

        y = priors[:, 1]
        x = priors[:, 0]
        x1 = (x - base_len * bbox_pred[:, 0]). \
            clamp(min=0, max=max_shape[1] - 1)
        y1 = (y - base_len * bbox_pred[:, 1]). \
            clamp(min=0, max=max_shape[0] - 1)
        x2 = (x + base_len * bbox_pred[:, 2]). \
            clamp(min=0, max=max_shape[1] - 1)
        y2 = (y + base_len * bbox_pred[:, 3]). \
            clamp(min=0, max=max_shape[0] - 1)
        decoded_bboxes = torch.stack([x1, y1, x2, y2], -1)
        return decoded_bboxes

    def _get_points_single(self, *args, **kwargs):
        """Get points according to feature map size.

        This function will be deprecated soon.
        """
        warnings.warn(
            '`_get_points_single` in `FoveaHead` will be '
            'deprecated soon, we support a multi level point generator now'
            'you can get points of a single level feature map '
            'with `self.prior_generator.single_level_grid_priors` ')
        y, x = super()._get_points_single(*args, **kwargs)
        return y + 0.5, x + 0.5