Paddle处理目标检测中的常用函数

氵文大师

已于 2022-12-20 16:04:04 修改

阅读量404

点赞数

分类专栏： paddlepaddle历险记文章标签：目标检测 paddle 深度学习

于 2022-12-07 01:30:44 首次发布

本文链接：https://blog.csdn.net/HaoZiHuang/article/details/128213305

版权

paddlepaddle历险记专栏收录该内容

46 篇文章 11 订阅

订阅专栏

1. $x_cy_cwh$ <=> $x_1y_1x_2y_2$

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 paddle.stack(b, axis=-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 paddle.stack(b, axis=-1)

输入 x shape 为 [num, 4]

2. 计算 $x_1y_1x_2y_2$ box的面积

def box_area(boxes: paddle.Tensor) -> paddle.Tensor:
    """
    Computes the area of a set of bounding boxes, which are specified by their
    (x1, y1, x2, y2) coordinates.

    Args:
        boxes (Tensor[N, 4]): boxes for which the area will be computed. They
            are expected to be in (x1, y1, x2, y2) format with
            ``0 <= x1 < x2`` and ``0 <= y1 < y2``.

    Returns:
        Tensor[N]: the area for each box
    """
    # boxes = _upcast(boxes)
    return (boxes[:, 2] - boxes[:, 0]) * (boxes[:, 3] - boxes[:, 1])

输入 x shape 为 [num, 4]

3. 计算 $x_1y_1x_2y_2$ box 的 IoU

def box_iou(boxes1, boxes2):
    area1 = box_area(boxes1)
    area2 = box_area(boxes2)

    lt = paddle.maximum(boxes1[:, None, :2], boxes2[:, :2])  # [N,M,2]
    rb = paddle.minimum(boxes1[:, None, 2:], boxes2[:, 2:])  # [N,M,2]

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

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

    iou = inter / union
    return iou, union

输入 boxes1 shape 为 [num1, 4]
输入 boxes2 shape 为 [num2, 4]
iou shape 为 [num1, num2]

4. 计算 $x_1y_1x_2y_2$ box 的 GIoU

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 = paddle.minimum(boxes1[:, None, :2], boxes2[:, :2])
    rb = paddle.maximum(boxes1[:, None, 2:], boxes2[:, 2:])

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

    return iou - (area - union) / area