根据预测图片和GT计算得到mPA和mIoU值

欧二lord

已于 2023-01-07 12:52:16 修改

阅读量679

点赞数

文章标签： python opencv 图像处理计算机视觉 pytorch

于 2022-06-12 15:49:10 首次发布

本文链接：https://blog.youkuaiyun.com/weixin_43117620/article/details/125246176

版权

这段代码实现了一个计算图像预测与Ground Truth之间的交并比（mIoU）和像素准确率（mPA）的功能。它通过遍历GT目录和预测目录中的PNG图像，使用fast_hist函数计算像素对应，并通过per_class_iu和per_class_PA函数计算每个类别的IoU和PA，最后输出平均值。适用于评估图像分割任务的性能。

摘要生成于 C知道，由 DeepSeek-R1 满血版支持，前往体验 >

如题，以下代码可以根据预测图片和GT得到mPA和mIoU的值。

注：使用此代码进行计算时，GT中的标签值需为从（0,0,0）递增至（n,n,n），n为类别的数量。

import numpy as np
from PIL import Image
from os.path import join


def fast_hist(a, b, n):
    k = (a >= 0) & (a < n)
    return np.bincount(n * a[k].astype(int) + b[k], minlength=n ** 2).reshape(n, n)


def per_class_iu(hist):
    return np.diag(hist) / np.maximum((hist.sum(1) + hist.sum(0) - np.diag(hist)), 1)


def per_class_PA(hist):
    return np.diag(hist) / np.maximum(hist.sum(1), 1)


def compute_mIoU(gt_dir, pred_dir, png_name_list, num_classes, name_classes):
    print('Num classes', num_classes)

    hist = np.zeros((num_classes, num_classes))

    gt_imgs = [join(gt_dir, x + ".png") for x in png_name_list]
    pred_imgs = [join(pred_dir, x + ".png") for x in png_name_list]

    for ind in range(len(gt_imgs)):

        pred = np.array(Image.open(pred_imgs[ind]))

        label = np.array(Image.open(gt_imgs[ind]))
        if len(label.flatten()) != len(pred.flatten()):
            print(
                'Skipping: len(gt) = {:d}, len(pred) = {:d}, {:s}, {:s}'.format(
                    len(label.flatten()), len(pred.flatten()), gt_imgs[ind],
                    pred_imgs[ind]))
            continue

        hist += fast_hist(label.flatten(), pred.flatten(), num_classes)

        if ind > 0 and ind % 10 == 0:
            print('{:d} / {:d}: mIou-{:0.2f}; mPA-{:0.2f}'.format(ind, len(gt_imgs),
                                                                  100 * np.nanmean(per_class_iu(hist)),
                                                                  100 * np.nanmean(per_class_PA(hist))))

    mIoUs = per_class_iu(hist)
    mPA = per_class_PA(hist)

    for ind_class in range(num_classes):
        print('===>' + name_classes[ind_class] + ':\tmIoU:' + str(round(mIoUs[ind_class] * 100, 2)) + '; mPA:' + str(
            round(mPA[ind_class] * 100, 2)))

    print('===> mIoU: ' + str(round(np.nanmean(mIoUs) * 100, 2)) + '; mPA: ' + str(round(np.nanmean(mPA) * 100, 2)))
    return mIoUs


png_name_list = [""]
name_classes = [""]
compute_mIoU("./gt", "./pred", png_name_list, 3, name_classes)