智能数字图像处理之FastRCNN(pytorch)代码解读之faster_rcnn_framework.py

class FasterRCNNBase(nn.Module):

广义R-CNN的主要类。参数:支柱(nn.Module):项(nn.Module):roi_heads (n . module):从RPN获取特性+建议并计算探测/遮罩。转换(n . module):执行从输入到feed的数据转换该模型 

 def __init__(self, backbone, rpn, roi_heads, transform):-》参数重新赋值給变量
        super(FasterRCNNBase, self).__init__()
        self.transform = transform
        self.backbone = backbone
        self.rpn = rpn
        self.roi_heads = roi_heads
        self._has_warned = False-》仅在torchscript模式下使用

    @torch.jit.unused
    def eager_outputs(self, losses, detections):-》训练时返回损失值测试时返回检测值
        # type: (Dict[str, Tensor], List[Dict[str, Tensor]]) -> Tuple[Dict[str, Tensor], List[Dict[str, Tensor]]]
        if self.training:
            return losses

        return detections

 def forward(self, images, targets=None):-》前向传播,参数:图像(列表[张量]):待处理的图像目标(list[Dict[张量]]):图像中出现的地面真值盒(可选)返回:结果(list[BoxList]或dict[张量]):模型的输出。在训练期间,它返回一个包含损失的dict[张量]。在测试期间,它返回包含其他字段的list[BoxList]比如“分数”、“标签”和“mask”(用于mask R-CNN模型)。
        if self.training and targets is None:-》传值为空报错
            raise ValueError("In training mode, targets should be passed")

        if self.training:
            assert targets is not None
            for target in targets:         # 进一步判断传入的target的boxes参数是否符合规定
                boxes = target["boxes"]-》取boxes值
                if isinstance(boxes, torch.Tensor):
                    if len(boxes.shape) != 2 or boxes.shape[-1] != 4:
                        raise ValueError("Expected target boxes to be a tensor"-》报错期望目标框是一个张量
                                         "of shape [N, 4], got {:}.".format(
                                          boxes.shape))
                else:
                    raise ValueError("Expected target boxes to be of type "-》报错预期的目标框是类型的
                                     "Tensor, got {:}.".format(type(boxes)))

        original_image_sizes = torch.jit.annotate(List[Tuple[int, int]], [])-》获取图像数据集大小
        for img in images:-》遍历图像
            val = img.shape[-2:]
            assert len(val) == 2  # 防止输入的是个一维向量
            original_image_sizes.append((val[0], val[1]))-》获取图像大小
        # original_image_sizes = [img.shape[-2:] for img in images]

        images, targets = self.transform(images, targets)  # 对图像进行预处理
        # print(images.tensors.shape)
        features = self.backbone(images.tensors)  # 将图像输入backbone得到特征图
        if isinstance(features, torch.Tensor):  # 若只在一层特征层上预测,将feature放入有序字典中,并编号为‘0’
            features = OrderedDict([('0', features)])  # 若在多层特征层上预测,传入的就是一个有序字典

        # 将特征层以及标注target信息传入rpn中
        proposals, proposal_losses = self.rpn(images, features, targets)

        # 将rpn生成的数据以及标注target信息传入fast rcnn后半部分
        detections, detector_losses = self.roi_heads(features, proposals, images.image_sizes, targets)

        # 对网络的预测结果进行后处理(主要将bboxes还原到原图像尺度上)
        detections = self.transform.postprocess(detections, images.image_sizes, original_image_sizes)

        losses = {}
        losses.update(detector_losses)-》更新检测损失值
        losses.update(proposal_losses)-》更新标签损失值

        if torch.jit.is_scripting():
            if not self._has_warned:
                warnings.warn("RCNN always returns a (Losses, Detections) tuple in scripting")-》输出警告信息
                self._has_warned = True
            return losses, detections
        else:
            return self.eager_outputs(losses, detections)
 

class TwoMLPHead(nn.Module):-》基于fpga的模型的标准头参数:in_channels (int):输入通道的数目representation_size (int):中间表示的大小

    def __init__(self, in_channels, representation_size):
        super(TwoMLPHead, self).__init__()

        self.fc6 = nn.Linear(in_channels, representation_size)
        self.fc7 = nn.Linear(representation_size, representation_size)

    def forward(self, x):
        x = x.flatten(start_dim=1)-》降维

        x = F.relu(self.fc6(x))-》降维
        x = F.relu(self.fc7(x))-》降维

        return x

class FastRCNNPredictor(nn.Module):-》标准分类+边界框回归层R-CNN为快。参数:in_channels (int):输入通道的数目
num_classes (int):输出类的数量(包括后台)

    def __init__(self, in_channels, num_classes):
        super(FastRCNNPredictor, self).__init__()
        self.cls_score = nn.Linear(in_channels, num_classes)
        self.bbox_pred = nn.Linear(in_channels, num_classes * 4)

    def forward(self, x):
        if x.dim() == 4:
            assert list(x.shape[2:]) == [1, 1]
        x = x.flatten(start_dim=1)
        scores = self.cls_score(x)-》全连接神经网络输出分类预测值
        bbox_deltas = self.bbox_pred(x)-》全连接神经网络输出目标框预测

        return scores, bbox_deltas

class FasterRCNN(FasterRCNNBase):

 def __init__(self, backbone, num_classes=None,
                 #转换参数
                 min_size=800, max_size=1000,      # 预处理resize时限制的最小尺寸与最大尺寸
                 image_mean=None, image_std=None,  # 预处理normalize时使用的均值和方差
                 # RPN 参数
                 rpn_anchor_generator=None, rpn_head=None,
                 rpn_pre_nms_top_n_train=2000, rpn_pre_nms_top_n_test=1000,    # rpn中在nms处理前保留的proposal数(根据score)
                 rpn_post_nms_top_n_train=2000, rpn_post_nms_top_n_test=1000,  # rpn中在nms处理后保留的proposal数
                 rpn_nms_thresh=0.7,  # rpn中进行nms处理时使用的iou阈值
                 rpn_fg_iou_thresh=0.7, rpn_bg_iou_thresh=0.3,  # rpn计算损失时,采集正负样本设置的阈值
                 rpn_batch_size_per_image=256, rpn_positive_fraction=0.5,  # rpn计算损失时采样的样本数,以及正样本占总样本的比例
                 # 框参数
                 box_roi_pool=None, box_head=None, box_predictor=None,
                 # 移除低目标概率      fast rcnn中进行nms处理的阈值   对预测结果根据score排序取前100个目标
                 box_score_thresh=0.05, box_nms_thresh=0.5, box_detections_per_img=100,
                 box_fg_iou_thresh=0.5, box_bg_iou_thresh=0.5,   # fast rcnn计算误差时,采集正负样本设置的阈值
                 box_batch_size_per_image=512, box_positive_fraction=0.25,  # fast rcnn计算误差时采样的样本数,以及正样本占所有样本的比例
                 bbox_reg_weights=None):
        if not hasattr(backbone, "out_channels"):
            raise ValueError(
                "backbone should contain an attribute out_channels"主干应该包含一个out_channels属性
                "specifying the number of output channels  (assumed to be the"指定输出通道的数量(假设为所有的级别都一样
                "same for all the levels"
            )

        assert isinstance(rpn_anchor_generator, (AnchorsGenerator, type(None)))-》断言
        assert isinstance(box_roi_pool, (MultiScaleRoIAlign, type(None)))

        if num_classes is not None:
            if box_predictor is not None:
                raise ValueError("num_classes should be None when box_predictor "当box_predictor时,num_classes应该为None
                                 "is specified")
        else:
            if box_predictor is None:
                raise ValueError("num_classes should not be None when box_predictor "
                                 "is not specified")

        # 预测特征层的channels
        out_channels = backbone.out_channels

        # 若anchor生成器为空,则自动生成针对resnet50_fpn的anchor生成器
        if rpn_anchor_generator is None:
            anchor_sizes = ((32,), (64,), (128,), (256,), (512,))
            aspect_ratios = ((0.5, 1.0, 2.0),) * len(anchor_sizes)
            rpn_anchor_generator = AnchorsGenerator(
                anchor_sizes, aspect_ratios
            )

        # 生成RPN通过滑动窗口预测网络部分
        if rpn_head is None:
            rpn_head = RPNHead(
                out_channels, rpn_anchor_generator.num_anchors_per_location()[0]
            )
        rpn_pre_nms_top_n = dict(training=rpn_pre_nms_top_n_train, testing=rpn_pre_nms_top_n_test)
        rpn_post_nms_top_n = dict(training=rpn_post_nms_top_n_train, testing=rpn_post_nms_top_n_test)

        # 定义整个RPN框架
        rpn = RegionProposalNetwork(
            rpn_anchor_generator, rpn_head,
            rpn_fg_iou_thresh, rpn_bg_iou_thresh,
            rpn_batch_size_per_image, rpn_positive_fraction,
            rpn_pre_nms_top_n, rpn_post_nms_top_n, rpn_nms_thresh)

        #  多尺度RoIAlign池化
        if box_roi_pool is None:
            box_roi_pool = MultiScaleRoIAlign(
                featmap_names=['0', '1', '2', '3'],  # 在哪些特征层进行roi pooling
                output_size=[7, 7],
                sampling_ratio=2)

        # fast RCNN中roi pooling后的展平处理两个全连接层部分
        if box_head is None:
            resolution = box_roi_pool.output_size[0]  # 默认等于7
            representation_size = 1024
            box_head = TwoMLPHead(
                out_channels * resolution ** 2,
                representation_size
            )

        # 在box_head的输出上预测部分
        if box_predictor is None:
            representation_size = 1024
            box_predictor = FastRCNNPredictor(
                representation_size,
                num_classes)

        # 将roi pooling, box_head以及box_predictor结合在一起
        roi_heads = RoIHeads(
            # box
            box_roi_pool, box_head, box_predictor,
            box_fg_iou_thresh, box_bg_iou_thresh,
            box_batch_size_per_image, box_positive_fraction,
            bbox_reg_weights,
            box_score_thresh, box_nms_thresh, box_detections_per_img)

        if image_mean is None:
            image_mean = [0.485, 0.456, 0.406]
        if image_std is None:
            image_std = [0.229, 0.224, 0.225]

        # 对数据进行标准化,缩放,打包成batch等处理部分
        transform = GeneralizedRCNNTransform(min_size, max_size, image_mean, image_std)

        super(FasterRCNN, self).__init__(backbone, rpn, roi_heads, transform)
 

import torch
from torch import nn
from collections import OrderedDict
from network_files.rpn_function import AnchorsGenerator, RPNHead, RegionProposalNetwork
from network_files.roi_head import RoIHeads
from torchvision.ops import MultiScaleRoIAlign
from torch.jit.annotations import Tuple, List, Dict, Optional
from torch import Tensor
import torch.nn.functional as F
import warnings
from network_files.transform import GeneralizedRCNNTransform


class FasterRCNNBase(nn.Module):
    """
    Main class for Generalized R-CNN.

    Arguments:
        backbone (nn.Module):
        rpn (nn.Module):
        roi_heads (nn.Module): takes the features + the proposals from the RPN and computes
            detections / masks from it.
        transform (nn.Module): performs the data transformation from the inputs to feed into
            the model
    """

    def __init__(self, backbone, rpn, roi_heads, transform):
        super(FasterRCNNBase, self).__init__()
        self.transform = transform
        self.backbone = backbone
        self.rpn = rpn
        self.roi_heads = roi_heads
        # used only on torchscript mode
        self._has_warned = False

    @torch.jit.unused
    def eager_outputs(self, losses, detections):
        # type: (Dict[str, Tensor], List[Dict[str, Tensor]]) -> Tuple[Dict[str, Tensor], List[Dict[str, Tensor]]]
        if self.training:
            return losses

        return detections

    def forward(self, images, targets=None):
        # type: (List[Tensor], Optional[List[Dict[str, Tensor]]]) -> Tuple[Dict[str, Tensor], List[Dict[str, Tensor]]]
        """
        Arguments:
            images (list[Tensor]): images to be processed
            targets (list[Dict[Tensor]]): ground-truth boxes present in the image (optional)

        Returns:
            result (list[BoxList] or dict[Tensor]): the output from the model.
                During training, it returns a dict[Tensor] which contains the losses.
                During testing, it returns list[BoxList] contains additional fields
                like `scores`, `labels` and `mask` (for Mask R-CNN models).

        """
        if self.training and targets is None:
            raise ValueError("In training mode, targets should be passed")

        if self.training:
            assert targets is not None
            for target in targets:         # 进一步判断传入的target的boxes参数是否符合规定
                boxes = target["boxes"]
                if isinstance(boxes, torch.Tensor):
                    if len(boxes.shape) != 2 or boxes.shape[-1] != 4:
                        raise ValueError("Expected target boxes to be a tensor"
                                         "of shape [N, 4], got {:}.".format(
                                          boxes.shape))
                else:
                    raise ValueError("Expected target boxes to be of type "
                                     "Tensor, got {:}.".format(type(boxes)))

        original_image_sizes = torch.jit.annotate(List[Tuple[int, int]], [])
        for img in images:
            val = img.shape[-2:]
            assert len(val) == 2  # 防止输入的是个一维向量
            original_image_sizes.append((val[0], val[1]))
        # original_image_sizes = [img.shape[-2:] for img in images]

        images, targets = self.transform(images, targets)  # 对图像进行预处理
        # print(images.tensors.shape)
        features = self.backbone(images.tensors)  # 将图像输入backbone得到特征图
        if isinstance(features, torch.Tensor):  # 若只在一层特征层上预测,将feature放入有序字典中,并编号为‘0’
            features = OrderedDict([('0', features)])  # 若在多层特征层上预测,传入的就是一个有序字典

        # 将特征层以及标注target信息传入rpn中
        proposals, proposal_losses = self.rpn(images, features, targets)

        # 将rpn生成的数据以及标注target信息传入fast rcnn后半部分
        detections, detector_losses = self.roi_heads(features, proposals, images.image_sizes, targets)

        # 对网络的预测结果进行后处理(主要将bboxes还原到原图像尺度上)
        detections = self.transform.postprocess(detections, images.image_sizes, original_image_sizes)

        losses = {}
        losses.update(detector_losses)
        losses.update(proposal_losses)

        if torch.jit.is_scripting():
            if not self._has_warned:
                warnings.warn("RCNN always returns a (Losses, Detections) tuple in scripting")
                self._has_warned = True
            return losses, detections
        else:
            return self.eager_outputs(losses, detections)

        # if self.training:
        #     return losses
        #
        # return detections


class TwoMLPHead(nn.Module):
    """
    Standard heads for FPN-based models

    Arguments:
        in_channels (int): number of input channels
        representation_size (int): size of the intermediate representation
    """

    def __init__(self, in_channels, representation_size):
        super(TwoMLPHead, self).__init__()

        self.fc6 = nn.Linear(in_channels, representation_size)
        self.fc7 = nn.Linear(representation_size, representation_size)

    def forward(self, x):
        x = x.flatten(start_dim=1)

        x = F.relu(self.fc6(x))
        x = F.relu(self.fc7(x))

        return x


class FastRCNNPredictor(nn.Module):
    """
    Standard classification + bounding box regression layers
    for Fast R-CNN.

    Arguments:
        in_channels (int): number of input channels
        num_classes (int): number of output classes (including background)
    """

    def __init__(self, in_channels, num_classes):
        super(FastRCNNPredictor, self).__init__()
        self.cls_score = nn.Linear(in_channels, num_classes)
        self.bbox_pred = nn.Linear(in_channels, num_classes * 4)

    def forward(self, x):
        if x.dim() == 4:
            assert list(x.shape[2:]) == [1, 1]
        x = x.flatten(start_dim=1)
        scores = self.cls_score(x)
        bbox_deltas = self.bbox_pred(x)

        return scores, bbox_deltas


class FasterRCNN(FasterRCNNBase):
    """
    Implements Faster R-CNN.

    The input to the model is expected to be a list of tensors, each of shape [C, H, W], one for each
    image, and should be in 0-1 range. Different images can have different sizes.

    The behavior of the model changes depending if it is in training or evaluation mode.

    During training, the model expects both the input tensors, as well as a targets (list of dictionary),
    containing:
        - boxes (FloatTensor[N, 4]): the ground-truth boxes in [x1, y1, x2, y2] format, with values
          between 0 and H and 0 and W
        - labels (Int64Tensor[N]): the class label for each ground-truth box

    The model returns a Dict[Tensor] during training, containing the classification and regression
    losses for both the RPN and the R-CNN.

    During inference, the model requires only the input tensors, and returns the post-processed
    predictions as a List[Dict[Tensor]], one for each input image. The fields of the Dict are as
    follows:
        - boxes (FloatTensor[N, 4]): the predicted boxes in [x1, y1, x2, y2] format, with values between
          0 and H and 0 and W
        - labels (Int64Tensor[N]): the predicted labels for each image
        - scores (Tensor[N]): the scores or each prediction

    Arguments:
        backbone (nn.Module): the network used to compute the features for the model.
            It should contain a out_channels attribute, which indicates the number of output
            channels that each feature map has (and it should be the same for all feature maps).
            The backbone should return a single Tensor or and OrderedDict[Tensor].
        num_classes (int): number of output classes of the model (including the background).
            If box_predictor is specified, num_classes should be None.
        min_size (int): minimum size of the image to be rescaled before feeding it to the backbone
        max_size (int): maximum size of the image to be rescaled before feeding it to the backbone
        image_mean (Tuple[float, float, float]): mean values used for input normalization.
            They are generally the mean values of the dataset on which the backbone has been trained
            on
        image_std (Tuple[float, float, float]): std values used for input normalization.
            They are generally the std values of the dataset on which the backbone has been trained on
        rpn_anchor_generator (AnchorGenerator): module that generates the anchors for a set of feature
            maps.
        rpn_head (nn.Module): module that computes the objectness and regression deltas from the RPN
        rpn_pre_nms_top_n_train (int): number of proposals to keep before applying NMS during training
        rpn_pre_nms_top_n_test (int): number of proposals to keep before applying NMS during testing
        rpn_post_nms_top_n_train (int): number of proposals to keep after applying NMS during training
        rpn_post_nms_top_n_test (int): number of proposals to keep after applying NMS during testing
        rpn_nms_thresh (float): NMS threshold used for postprocessing the RPN proposals
        rpn_fg_iou_thresh (float): minimum IoU between the anchor and the GT box so that they can be
            considered as positive during training of the RPN.
        rpn_bg_iou_thresh (float): maximum IoU between the anchor and the GT box so that they can be
            considered as negative during training of the RPN.
        rpn_batch_size_per_image (int): number of anchors that are sampled during training of the RPN
            for computing the loss
        rpn_positive_fraction (float): proportion of positive anchors in a mini-batch during training
            of the RPN
        box_roi_pool (MultiScaleRoIAlign): the module which crops and resizes the feature maps in
            the locations indicated by the bounding boxes
        box_head (nn.Module): module that takes the cropped feature maps as input
        box_predictor (nn.Module): module that takes the output of box_head and returns the
            classification logits and box regression deltas.
        box_score_thresh (float): during inference, only return proposals with a classification score
            greater than box_score_thresh
        box_nms_thresh (float): NMS threshold for the prediction head. Used during inference
        box_detections_per_img (int): maximum number of detections per image, for all classes.
        box_fg_iou_thresh (float): minimum IoU between the proposals and the GT box so that they can be
            considered as positive during training of the classification head
        box_bg_iou_thresh (float): maximum IoU between the proposals and the GT box so that they can be
            considered as negative during training of the classification head
        box_batch_size_per_image (int): number of proposals that are sampled during training of the
            classification head
        box_positive_fraction (float): proportion of positive proposals in a mini-batch during training
            of the classification head
        bbox_reg_weights (Tuple[float, float, float, float]): weights for the encoding/decoding of the
            bounding boxes

    """

    def __init__(self, backbone, num_classes=None,
                 # transform parameter
                 min_size=800, max_size=1000,      # 预处理resize时限制的最小尺寸与最大尺寸
                 image_mean=None, image_std=None,  # 预处理normalize时使用的均值和方差
                 # RPN parameters
                 rpn_anchor_generator=None, rpn_head=None,
                 rpn_pre_nms_top_n_train=2000, rpn_pre_nms_top_n_test=1000,    # rpn中在nms处理前保留的proposal数(根据score)
                 rpn_post_nms_top_n_train=2000, rpn_post_nms_top_n_test=1000,  # rpn中在nms处理后保留的proposal数
                 rpn_nms_thresh=0.7,  # rpn中进行nms处理时使用的iou阈值
                 rpn_fg_iou_thresh=0.7, rpn_bg_iou_thresh=0.3,  # rpn计算损失时,采集正负样本设置的阈值
                 rpn_batch_size_per_image=256, rpn_positive_fraction=0.5,  # rpn计算损失时采样的样本数,以及正样本占总样本的比例
                 # Box parameters
                 box_roi_pool=None, box_head=None, box_predictor=None,
                 # 移除低目标概率      fast rcnn中进行nms处理的阈值   对预测结果根据score排序取前100个目标
                 box_score_thresh=0.05, box_nms_thresh=0.5, box_detections_per_img=100,
                 box_fg_iou_thresh=0.5, box_bg_iou_thresh=0.5,   # fast rcnn计算误差时,采集正负样本设置的阈值
                 box_batch_size_per_image=512, box_positive_fraction=0.25,  # fast rcnn计算误差时采样的样本数,以及正样本占所有样本的比例
                 bbox_reg_weights=None):
        if not hasattr(backbone, "out_channels"):
            raise ValueError(
                "backbone should contain an attribute out_channels"
                "specifying the number of output channels  (assumed to be the"
                "same for all the levels"
            )

        assert isinstance(rpn_anchor_generator, (AnchorsGenerator, type(None)))
        assert isinstance(box_roi_pool, (MultiScaleRoIAlign, type(None)))

        if num_classes is not None:
            if box_predictor is not None:
                raise ValueError("num_classes should be None when box_predictor "
                                 "is specified")
        else:
            if box_predictor is None:
                raise ValueError("num_classes should not be None when box_predictor "
                                 "is not specified")

        # 预测特征层的channels
        out_channels = backbone.out_channels

        # 若anchor生成器为空,则自动生成针对resnet50_fpn的anchor生成器
        if rpn_anchor_generator is None:
            anchor_sizes = ((32,), (64,), (128,), (256,), (512,))
            aspect_ratios = ((0.5, 1.0, 2.0),) * len(anchor_sizes)
            rpn_anchor_generator = AnchorsGenerator(
                anchor_sizes, aspect_ratios
            )

        # 生成RPN通过滑动窗口预测网络部分
        if rpn_head is None:
            rpn_head = RPNHead(
                out_channels, rpn_anchor_generator.num_anchors_per_location()[0]
            )

        # 默认rpn_pre_nms_top_n_train = 2000, rpn_pre_nms_top_n_test = 1000,
        # 默认rpn_post_nms_top_n_train = 2000, rpn_post_nms_top_n_test = 1000,
        rpn_pre_nms_top_n = dict(training=rpn_pre_nms_top_n_train, testing=rpn_pre_nms_top_n_test)
        rpn_post_nms_top_n = dict(training=rpn_post_nms_top_n_train, testing=rpn_post_nms_top_n_test)

        # 定义整个RPN框架
        rpn = RegionProposalNetwork(
            rpn_anchor_generator, rpn_head,
            rpn_fg_iou_thresh, rpn_bg_iou_thresh,
            rpn_batch_size_per_image, rpn_positive_fraction,
            rpn_pre_nms_top_n, rpn_post_nms_top_n, rpn_nms_thresh)

        #  Multi-scale RoIAlign pooling
        if box_roi_pool is None:
            box_roi_pool = MultiScaleRoIAlign(
                featmap_names=['0', '1', '2', '3'],  # 在哪些特征层进行roi pooling
                output_size=[7, 7],
                sampling_ratio=2)

        # fast RCNN中roi pooling后的展平处理两个全连接层部分
        if box_head is None:
            resolution = box_roi_pool.output_size[0]  # 默认等于7
            representation_size = 1024
            box_head = TwoMLPHead(
                out_channels * resolution ** 2,
                representation_size
            )

        # 在box_head的输出上预测部分
        if box_predictor is None:
            representation_size = 1024
            box_predictor = FastRCNNPredictor(
                representation_size,
                num_classes)

        # 将roi pooling, box_head以及box_predictor结合在一起
        roi_heads = RoIHeads(
            # box
            box_roi_pool, box_head, box_predictor,
            box_fg_iou_thresh, box_bg_iou_thresh,
            box_batch_size_per_image, box_positive_fraction,
            bbox_reg_weights,
            box_score_thresh, box_nms_thresh, box_detections_per_img)

        if image_mean is None:
            image_mean = [0.485, 0.456, 0.406]
        if image_std is None:
            image_std = [0.229, 0.224, 0.225]

        # 对数据进行标准化,缩放,打包成batch等处理部分
        transform = GeneralizedRCNNTransform(min_size, max_size, image_mean, image_std)

        super(FasterRCNN, self).__init__(backbone, rpn, roi_heads, transform)

https://github.com/WZMIAOMIAO/deep-learning-for-image-processing

©️2020 CSDN 皮肤主题: 编程工作室 设计师:CSDN官方博客 返回首页