RT-ODLab/utils/box_ops.py

import torch
import math
import numpy as np
from torchvision.ops.boxes import box_area


# ------------------ Box ops ------------------
def box_cxcywh_to_xyxy(x):
    x_c, y_c, w, h = x.unbind(-1)
    b = [(x_c - 0.5 * w), (y_c - 0.5 * h),
         (x_c + 0.5 * w), (y_c + 0.5 * h)]
    return torch.stack(b, dim=-1)

def box_xyxy_to_cxcywh(x):
    x0, y0, x1, y1 = x.unbind(-1)
    b = [(x0 + x1) / 2, (y0 + y1) / 2,
         (x1 - x0), (y1 - y0)]
    return torch.stack(b, dim=-1)

def rescale_bboxes(bboxes, origin_img_size, cur_img_size, deltas=None):
    origin_h, origin_w = origin_img_size
    cur_img_h, cur_img_w = cur_img_size
    if deltas is None:
        # rescale
        bboxes[..., [0, 2]] = bboxes[..., [0, 2]] / cur_img_w * origin_w
        bboxes[..., [1, 3]] = bboxes[..., [1, 3]] / cur_img_h * origin_h

        # clip bboxes
        bboxes[..., [0, 2]] = np.clip(bboxes[..., [0, 2]], a_min=0., a_max=origin_w)
        bboxes[..., [1, 3]] = np.clip(bboxes[..., [1, 3]], a_min=0., a_max=origin_h)
    else:
        # rescale
        bboxes[..., [0, 2]] = bboxes[..., [0, 2]] / (cur_img_w - deltas[0]) * origin_w
        bboxes[..., [1, 3]] = bboxes[..., [1, 3]] / (cur_img_h - deltas[1]) * origin_h
        
        # clip bboxes
        bboxes[..., [0, 2]] = np.clip(bboxes[..., [0, 2]], a_min=0., a_max=origin_w)
        bboxes[..., [1, 3]] = np.clip(bboxes[..., [1, 3]], a_min=0., a_max=origin_h)

    return bboxes

def bbox2dist(anchor_points, bbox, reg_max):
    '''Transform bbox(xyxy) to dist(ltrb).'''
    x1y1, x2y2 = torch.split(bbox, 2, -1)
    lt = anchor_points - x1y1
    rb = x2y2 - anchor_points
    dist = torch.cat([lt, rb], -1).clamp(0, reg_max - 0.01)
    return dist


# ------------------ IoU ops ------------------
def box_iou(boxes1, boxes2):
    area1 = box_area(boxes1)
    area2 = box_area(boxes2)

    lt = torch.max(boxes1[:, None, :2], boxes2[:, :2])  # [N,M,2]
    rb = torch.min(boxes1[:, None, 2:], boxes2[:, 2:])  # [N,M,2]

    wh = (rb - lt).clamp(min=0)  # [N,M,2]
    inter = wh[:, :, 0] * wh[:, :, 1]  # [N,M]

    union = area1[:, None] + area2 - inter

    iou = inter / union
    return iou, union

def generalized_box_iou(boxes1, boxes2):
    """
    Generalized IoU from https://giou.stanford.edu/

    The boxes should be in [x0, y0, x1, y1] format

    Returns a [N, M] pairwise matrix, where N = len(boxes1)
    and M = len(boxes2)
    """
    # degenerate boxes gives inf / nan results
    # so do an early check
    assert (boxes1[:, 2:] >= boxes1[:, :2]).all()
    assert (boxes2[:, 2:] >= boxes2[:, :2]).all()
    iou, union = box_iou(boxes1, boxes2)

    lt = torch.min(boxes1[:, None, :2], boxes2[:, :2])
    rb = torch.max(boxes1[:, None, 2:], boxes2[:, 2:])

    wh = (rb - lt).clamp(min=0)  # [N,M,2]
    area = wh[:, :, 0] * wh[:, :, 1]

    return iou - (area - union) / area

def get_ious(bboxes1,
             bboxes2,
             box_mode="xyxy",
             iou_type="iou"):
    """
    Compute iou loss of type ['iou', 'giou', 'linear_iou']

    Args:
        inputs (tensor): pred values
        targets (tensor): target values
        weight (tensor): loss weight
        box_mode (str): 'xyxy' or 'ltrb', 'ltrb' is currently supported.
        loss_type (str): 'giou' or 'iou' or 'linear_iou'
        reduction (str): reduction manner

    Returns:
        loss (tensor): computed iou loss.
    """
    if box_mode == "ltrb":
        bboxes1 = torch.cat((-bboxes1[..., :2], bboxes1[..., 2:]), dim=-1)
        bboxes2 = torch.cat((-bboxes2[..., :2], bboxes2[..., 2:]), dim=-1)
    elif box_mode != "xyxy":
        raise NotImplementedError

    eps = torch.finfo(torch.float32).eps

    bboxes1_area = (bboxes1[..., 2] - bboxes1[..., 0]).clamp_(min=0) \
        * (bboxes1[..., 3] - bboxes1[..., 1]).clamp_(min=0)
    bboxes2_area = (bboxes2[..., 2] - bboxes2[..., 0]).clamp_(min=0) \
        * (bboxes2[..., 3] - bboxes2[..., 1]).clamp_(min=0)

    w_intersect = (torch.min(bboxes1[..., 2], bboxes2[..., 2])
                   - torch.max(bboxes1[..., 0], bboxes2[..., 0])).clamp_(min=0)
    h_intersect = (torch.min(bboxes1[..., 3], bboxes2[..., 3])
                   - torch.max(bboxes1[..., 1], bboxes2[..., 1])).clamp_(min=0)

    area_intersect = w_intersect * h_intersect
    area_union = bboxes2_area + bboxes1_area - area_intersect
    ious = area_intersect / area_union.clamp(min=eps)

    if iou_type == "iou":
        return ious
    elif iou_type == "giou":
        g_w_intersect = torch.max(bboxes1[..., 2], bboxes2[..., 2]) \
            - torch.min(bboxes1[..., 0], bboxes2[..., 0])
        g_h_intersect = torch.max(bboxes1[..., 3], bboxes2[..., 3]) \
            - torch.min(bboxes1[..., 1], bboxes2[..., 1])
        ac_uion = g_w_intersect * g_h_intersect
        gious = ious - (ac_uion - area_union) / ac_uion.clamp(min=eps)
        return gious
    else:
        raise NotImplementedError

# copy from YOLOv5
def bbox_iou(box1, box2, xywh=True, GIoU=False, DIoU=False, CIoU=False, eps=1e-7):
    # Returns Intersection over Union (IoU) of box1(1,4) to box2(n,4)

    # Get the coordinates of bounding boxes
    if xywh:  # transform from xywh to xyxy
        (x1, y1, w1, h1), (x2, y2, w2, h2) = box1.chunk(4, -1), box2.chunk(4, -1)
        w1_, h1_, w2_, h2_ = w1 / 2, h1 / 2, w2 / 2, h2 / 2
        b1_x1, b1_x2, b1_y1, b1_y2 = x1 - w1_, x1 + w1_, y1 - h1_, y1 + h1_
        b2_x1, b2_x2, b2_y1, b2_y2 = x2 - w2_, x2 + w2_, y2 - h2_, y2 + h2_
    else:  # x1, y1, x2, y2 = box1
        b1_x1, b1_y1, b1_x2, b1_y2 = box1.chunk(4, -1)
        b2_x1, b2_y1, b2_x2, b2_y2 = box2.chunk(4, -1)
        w1, h1 = b1_x2 - b1_x1, b1_y2 - b1_y1 + eps
        w2, h2 = b2_x2 - b2_x1, b2_y2 - b2_y1 + eps

    # Intersection area
    inter = (b1_x2.minimum(b2_x2) - b1_x1.maximum(b2_x1)).clamp(0) * \
            (b1_y2.minimum(b2_y2) - b1_y1.maximum(b2_y1)).clamp(0)

    # Union Area
    union = w1 * h1 + w2 * h2 - inter + eps

    # IoU
    iou = inter / union
    if CIoU or DIoU or GIoU:
        cw = b1_x2.maximum(b2_x2) - b1_x1.minimum(b2_x1)  # convex (smallest enclosing box) width
        ch = b1_y2.maximum(b2_y2) - b1_y1.minimum(b2_y1)  # convex height
        if CIoU or DIoU:  # Distance or Complete IoU https://arxiv.org/abs/1911.08287v1
            c2 = cw ** 2 + ch ** 2 + eps  # convex diagonal squared
            rho2 = ((b2_x1 + b2_x2 - b1_x1 - b1_x2) ** 2 + (b2_y1 + b2_y2 - b1_y1 - b1_y2) ** 2) / 4  # center dist ** 2
            if CIoU:  # https://github.com/Zzh-tju/DIoU-SSD-pytorch/blob/master/utils/box/box_utils.py#L47
                v = (4 / math.pi ** 2) * (torch.atan(w2 / h2) - torch.atan(w1 / h1)).pow(2)
                with torch.no_grad():
                    alpha = v / (v - iou + (1 + eps))
                return iou - (rho2 / c2 + v * alpha)  # CIoU
            return iou - rho2 / c2  # DIoU
        c_area = cw * ch + eps  # convex area
        return iou - (c_area - union) / c_area  # GIoU https://arxiv.org/pdf/1902.09630.pdf
    return iou  # IoU


if __name__ == '__main__':
    box1 = torch.tensor([[10, 10, 20, 20]])
    box2 = torch.tensor([[15, 15, 20, 20]])
    iou = box_iou(box1, box2)
    print(iou)