YOLO坐标转真实坐标/YOLO与VOC标注转化

一、标签定义

YOLO标签

object-class> <x> <y> <width> <height>
0 0.371341 0.546939 0.474576 0.885714

x,y是目标的中心坐标，width,height是目标的宽和高。这些坐标是通过归一化的，其中x，width是使用原图的width进行归一化；而y，height是使用原图的height进行归一化。

VOC标签

<xmin>241</xmin>
<ymin>268</ymin>
<xmax>308</xmax>
<ymax>434</ymax>

(xmin,ymin)是左上角坐标,(xmax,ymax)是右下角坐标。

二、代码

yolo_to_voc

def yolo_to_voc(size, box):
    x = box[0] * size[0]
    w = box[2] * size[0]
    y = box[1] * size[1]
    h = box[3] * size[1]
    xmin = int(x - w/2)
    xmax = int(x + w/2)
    ymin = int(y - h/2)
    ymax = int(y + h/2)
    return (xmin, ymin, xmax, ymax)
 
size=(649,490)
box=[0.371341, 0.546939, 0.474576, 0.885714]
print(yolo_to_voc(size,box))

(87, 51, 395, 485)

voc_to_yolo

def voc_to_yolo(size, box):
    dw = 1./size[0]
    dh = 1./size[1]
    x = (box[0] + box[2])/2.0
    y = (box[1] + box[3])/2.0
    w = box[2] - box[0]
    h = box[3] - box[1]
    x = x*dw
    w = w*dw
    y = y*dh
    h = h*dh
    return (x,y,w,h)

size=(649,490)
box=[87, 51, 395, 485]
print(voc_to_yolo(size,box))

(0.3713405238828968, 0.5469387755102041, 0.4745762711864407, 0.8857142857142858)

原文链接：https://blog.csdn.net/KD_LW/article/details/112918173
原文中yolo_to_voc是错误的