DLIB：训练 Shape_predictor 194 个地标（海伦数据集）

Question

我正在使用 helen 数据集[=训练 DLIB 的 shape_predictor 的 194 个面部特征点 用于通过 dlib 库的 face_landmark_detection_ex.cpp 检测人脸标志。

现在它给了我一个 sp.dat 大约 45 MB 的二进制文件，与给定的文件 (http://sourceforge.net/projects/dclib/files/dlib/v18.10/shape_predictor_68_face_landmarks.dat.bz2) 相比，它有 68 个面部特征点。训练中

平均训练误差：0.0203811
平均测试误差：0.0204511

当我使用经过训练的数据来获取面部标志位置时，我得到的结果是..

这与从 68 个地标得到的结果有很大偏差

68 张地标图片：

为什么？

Answer 1

好的，看来您还没有阅读 code 评论 (?):

shape_predictor_trainer trainer;
// This algorithm has a bunch of parameters you can mess with.  The
// documentation for the shape_predictor_trainer explains all of them.
// You should also read Kazemi's paper which explains all the parameters
// in great detail.  However, here I'm just setting three of them
// differently than their default values.  I'm doing this because we
// have a very small dataset.  In particular, setting the oversampling
// to a high amount (300) effectively boosts the training set size, so
// that helps this example.
trainer.set_oversampling_amount(300);
// I'm also reducing the capacity of the model by explicitly increasing
// the regularization (making nu smaller) and by using trees with
// smaller depths.  
trainer.set_nu(0.05);
trainer.set_tree_depth(2);

查看 Kazemi paper，ctrl-f 字符串 'parameter' 并阅读...

DLIB：训练 Shape_predictor 194 个地标（海伦数据集）

DLIB : Training Shape_predictor for 194 landmarks (helen dataset)

c++

feature-detection

training-data

dlib