人脸检测——Faceboxes

studyeboy

于 2023-03-23 16:35:12 发布

阅读量117

点赞数

分类专栏：算法图像处理 pytorch 文章标签： python pytorch 深度学习

本文链接：https://blog.csdn.net/studyeboy/article/details/129733528

版权

算法同时被 3 个专栏收录

47 篇文章 3 订阅

订阅专栏

图像处理

41 篇文章 10 订阅

订阅专栏

pytorch

15 篇文章 1 订阅

订阅专栏

FaceBoxes.PyTorch

安装
```
git clone https://github.com/zisianw/FaceBoxes.PyTorch.git
sh ./make.sh
```
我的系统中使用的Python3.8，直接sh ./make.sh使用的是python3.6，所以修改make.sh为：
模型文件
需要用到的文件，nms中的编译文件为utils->nms中的编译后文件。
使用

#faceboxes detector initialize
def faceboxes_init(model_dir):
    net = FaceBoxes(phase='test', size=None, num_classes=2)
    net.load_state_dict(torch.load(model_dir))
    cuda = torch.cuda.is_available()
    if cuda:
        net.cuda()

    net.eval()

    return net, cuda

def faceboxes_dectect(model, cuda, image):
    img = np.float32(image)
    h, w = img.shape[:2]
    scale = torch.Tensor([img.shape[1], img.shape[0], img.shape[1], img.shape[0]])
    img -= (104, 117, 123)
    img = img.transpose(2, 0, 1)
    img = torch.from_numpy(img).unsqueeze(0)
    if cuda:
        img = img.cuda()
        scale = scale.cuda()

    loc, conf = model(img)
    priorbox = PriorBox(cfg, image_size=(h, w))
    priors = priorbox.forward()
    if cuda:
        priors = priors.cuda()
    prior_data = priors.data
    boxes = decode(loc.data.squeeze(0), prior_data, cfg['variance'])
    boxes = boxes * scale
    boxes = boxes.cpu().numpy()
    scores = conf.squeeze(0).data.cpu().numpy()[:, 1]

    #ignore low scores
    inds = np.where(scores > 0.7)[0]
    boxes = boxes[inds]
    scores = scores[inds]

    #keep top-K before NMS
    order = scores.argsort()[::-1][:5000]
    boxes = boxes[order]
    scores = scores[order]

    #do NMS
    dets = np.hstack((boxes, scores[:, np.newaxis])).astype(np.float32, copy=False)
    keep = nms(dets, 0.3, force_cpu=False)
    dets = dets[keep, :]

    #keep top-K faster NMS
    dets = dets[: 750, :]

    return dets


faceboxes_model, faceboxes_cuda = faceboxes_init(args.faceboxes_detect_model_dir)
result_bboxes = faceboxes_dectect(faceboxes_model, faceboxes_cuda, image)