def annotations_to_instances()

in meshrcnn/data/meshrcnn_transforms.py [0:0]

• 24 lines of code
• 18 McCabe index (conditional complexity)

def annotations_to_instances(annos, image_size):
    """
    Create an :class:`Instances` object used by the models,
    from instance annotations in the dataset dict.

    Args:
        annos (list[dict]): a list of annotations, one per instance.
        image_size (tuple): height, width

    Returns:
        Instances: It will contains fields "gt_boxes", "gt_classes",
            "gt_masks", "gt_keypoints", if they can be obtained from `annos`.
    """
    boxes = [BoxMode.convert(obj["bbox"], obj["bbox_mode"], BoxMode.XYXY_ABS) for obj in annos]
    target = Instances(image_size)
    boxes = target.gt_boxes = Boxes(boxes)
    boxes.clip(image_size)

    classes = [obj["category_id"] for obj in annos]
    classes = torch.tensor(classes, dtype=torch.int64)
    target.gt_classes = classes

    if len(annos) and "segmentation" in annos[0]:
        masks = [obj["segmentation"] for obj in annos]
        target.gt_masks = torch.stack(masks, dim=0)

    # camera
    if len(annos) and "K" in annos[0]:
        K = [torch.tensor(obj["K"]) for obj in annos]
        target.gt_K = torch.stack(K, dim=0)

    if len(annos) and "voxel" in annos[0]:
        voxels = [obj["voxel"] for obj in annos]
        target.gt_voxels = VoxelInstances(voxels)

    if len(annos) and "mesh" in annos[0]:
        meshes = [obj["mesh"] for obj in annos]
        target.gt_meshes = MeshInstances(meshes)

    if len(annos) and "dz" in annos[0]:
        dz = [obj["dz"] for obj in annos]
        target.gt_dz = torch.tensor(dz)

    return target