Preprocess an image. 1. rescale the image to expected size 2. normalize the image 3. flip the image (if needed) 4. pad the image (if

Preprocess gt bboxes. 1. rescale bboxes according to image size 2. flip bboxes (if needed) 3. pad the first dimension to `max_num_gts`

A wrapper of concatenated dataset. Same as :obj:`torch.utils.data.dataset.ConcatDataset`, but concat the group flag for image aspect ratio.

Sampler that restricts data loading to a subset of the dataset. It is especially useful in conjunction with :class:`torch.nn.parallel.Distribu

Preprocess masks. 1. resize masks to expected size and stack to a single array 2. flip the masks (if needed) 3. pad the masks (if needed)

Anchor-based head (RPN, RetinaNet, SSD, etc.). Args: in_channels (int): Number of channels in the input feature map. feat_channel

Extract RoI features from all level feature map and concat them for consistent supervison. Args: roi_layer (dict): Specify RoI layer type

Simplest RoI head, with only two fc layers for classification and regression respectively

More general bbox head, with shared conv and fc layers and two optional separated branches. /-> cls convs -> cls

Simplest RoI head, with only two fc layers for classification and regression respectively

Base class for detectors

More general bbox head, with shared conv and fc layers and two optional separated branches. /-> cls convs -> cls

Custom dataset for detection. Annotation format: [ { 'filename': 'a.jpg', 'width': 1280, 'height'

Assign a corresponding gt bbox or background to each bbox. Each proposals will be assigned with `-1`, `0`, or a positive integer indicating t

ResNeXt backbone. Args: depth (int): Depth of resnet, from {18, 34, 50, 101, 152}. num_stages (int): Resnet stages, normally 4.

ResNet backbone. Args: depth (int): Depth of resnet, from {18, 34, 50, 101, 152}. num_stages (int): Resnet stages, normally 4.

Extract RoI features from a single level feature map. If there are mulitple input feature levels, each RoI is mapped to a level according to

Extract RoI features from all level feature map. If there are mulitple input feature levels, each RoI is mapped to a level according to it