PFLD-pytorch

PFLD-pytorch copied to clipboard

Bumps [opencv-python](https://github.com/skvark/opencv-python) from 4.1.0.25 to 4.2.0.32. Release notes Sourced from opencv-python's releases. 4.2.0.32 opencv-python: https://pypi.org/project/opencv-python/ opencv-contrib-python: https://pypi.org/project/opencv-contrib-python/ opencv-python-headless: https://pypi.org/project/opencv-python-headless/ opencv-contrib-python-headless: https://pypi.org/project/opencv-contrib-python-headless/ OpenCV version 4.2.0. Changes: macOS environment updated from xcode8.3...

dependabot[bot]

在train.py文件中， pic should be PIL Image or ndarray. Got <class 'NoneType>该如何解决这个问题呢

4

yangazhang

I am testing your model and the result looks pretty weird, some values are more than 1, so when multiply by 112, the value is out of the range [0,...

HariWu1995

执行pytorch2onnx.py时报错

1

您好，我在执行pytorch2onnx.py这个文件的时候报错： ====> check onnx model... ====> Simplifying... Traceback (most recent call last): File "pytorch2onnx.py", line 46, in onnx.save(model_opt, args.onnx_model_sim) File "C:\Users\Administrator\anaconda3\envs\py37\lib\site-packages\onnx\__init__.py", line 184, in save_model proto = write_external_data_tensors(proto, basepath) File...

ZZHHogan

测试效果差（距离镜头近）。

8

感谢分享代码。我用自己拍摄的人脸图，通过MTCNN检测人脸，再进行pfld关键点预测，测试效果非常不好。造成这个的原因是pfld模型设计本身的原因？还是训练数据集和测试数据差别太大(训练数据集都是一些裁剪出来的低分辨率小图[112*112]，测试的时候是用手机拍摄的高分辨率大图)？通过对这种自己拍摄的高清大图进行手动的人脸裁剪再送入pfld进行关键点检测可以一定程度上提高测试效果，但是针对不同的图没有一个相同的裁剪标准。。。。。想问一下在人脸检测的基础上抠人脸图应遵循一个什么标准。

liuxx2016

euler angles

2

Hello! https://github.com/polarisZhao/PFLD-pytorch/blob/e438e2bae50cfb12e3c2ee6532e906c84f5660fc/pfld/utils.py#L29 here it is said about axes, which were used as basis for euler angle calculation, but it is not clear for me as on https://github.com/polarisZhao/PFLD-pytorch/blob/e438e2bae50cfb12e3c2ee6532e906c84f5660fc/pfld/utils.py#L32 it is said...

kurbobo

The training set does not contain IBUG, and IBUG belongs to the validation set. 

iam-zhanghongliang

在CPU上测试很慢

2

直接测试每秒大概5帧左右，有什么建议和指导吗？

NGA-MMDDsteven

网络前几层的bias为何置为False

1

rt

lwihtittlee

I used 3D tools to annotate the dataset for 98 landmarks. Attaching the sample images annotated using 3D Tool. But when I did Transfer Learning on the same for 350...

Kanika-AI-Learning-Repo