FasterRCNN源码解析（九）——FastRCNN正负样本划分及采样

FasterRCNN源码解析（九）——FastRCNN正负样本划分及采样

对应框架图中的这个部分

select_training_samples

主要是针对FasterRCNN源码解析（八）——ROIheads部分roi_head正向传播过程中第一步self.select_training_samples(proposals, targets)进行一个详解，如何对proposal划分正负样本，统计对应gt的标签以及边界框回归信息

proposals, matched_idxs, labels, regression_targets = self.select_training_samples(proposals, targets)

    def select_training_samples(self,
                                proposals,  # type: List[Tensor]
                                targets     # type: Optional[List[Dict[str, Tensor]]]
                                ):
        # type: (...) -> Tuple[List[Tensor], List[Tensor], List[Tensor], List[Tensor]]
        """
        划分正负样本，统计对应gt的标签以及边界框回归信息
        list元素个数为batch_size
        Args:
            proposals: rpn预测的boxes
            targets:

        Returns:

        """

        # 检查target数据是否为空
        self.check_targets(targets)
        assert targets is not None
        dtype = proposals[0].dtype
        device = proposals[0].device

        # 获取标注好的boxes 以及 labels 信息
        gt_boxes = [t["boxes"].to(dtype) for t in targets]
        gt_labels = [t["labels"] for t in targets]

        # append ground-truth bboxes to proposal
        # 将gt_boxes拼接到proposal后面
        proposals = self.add_gt_proposals(proposals