-
您好,我现在手上有106点数据集,在wflw数据集的数据增强的时候用到了Mirror98.TXT,这个文档中记录的点顺序有什么规律呢,我用106点数据集该怎么准备这个txt呢。谢谢。
-
I have tried both with "pretrained_model/mobileNetV2_0.25.pth" and without this pretrained model for retraining the model for common, challenge and fullset.
However, losses never goes down to below…
-
I am curious to know what kind of DL inference framework used this HyperLandmark android app?
Can you share this information?
-
Hello,thank you ao much for awesome work,how yo get landmarks points from this model?
Im using Face recognition and i need to have facial landmarks.
-
得到的onnx_output.jpg的显示结果全都在图像的左上角, 而且也不像是等比例缩放的结果,这是为什么呢?
-
The paper mentioned that the method first trains on wflw dataset and use the weights as pretrain model and train on 300w
However, 300w just has 68 points without angle. How the method work with th…
-
Step1: Prepare the training data
Align the face images to 112*112 according to face_align.py.
需要用face_align.py这个文件处理图片还是简简单单把图片resize成112*112呢
-
你好,请问如果我的关键点检测模型是接在人脸检测模型后面,但是我在训练过程中,保证了所有关键点均在图片上,但是在与检测模型进行结合的时候,送入的框并不包含所有的关键点,但是对于关键点检测模型而言,还是认为点在框里面,这样进行检测会使得点进行偏移,预测的就有问题,请问这个应该如何解决?如果方便的话,麻烦加一下我的qq:327595697,想找个人探讨一下,谢谢
-
lz,我用其他的自己拍的人脸图,发现人脸占比和训练数据的不同,因此就效果不好。在你的视频测试里,是先用人脸检测的网络,来把人脸扣出来,然后再给pfld检测,抠图的比例是在检测框的基础上取长宽最大值的1.1倍。那么我自己抠的人脸图,是要遵循什么样的人脸占比来才好呢?还是我用其他的人脸检测网络,把检测网络的抠图范围扩大1.1倍就都可以用呢?
-
RuntimeError: ./checkpoint/snapshot/checkpoint.pth.tar is a zip archive (did you mean to use torch.jit.load()?)