Please wait a minute...
FITEE
Front. Inform. Technol. Electron. Eng.  2015, Vol. 16 Issue (11): 930-939    DOI: 10.1631/FITEE.1500125
    
基于深度卷积网络的多分类法在头部姿态估计中的应用
Ying Cai, Meng-long Yang, Jun Li
School of Computer Science, Sichuan University, Chengdu 610065, China; College of Information Engineering, Sichuan Agricultural University, Yaan 625014, China; School of Aeronautics and Astronautics, Sichuan University, Chengdu 610065, China
Multiclass classification based on a deep convolutional network for head pose estimation
Ying Cai, Meng-long Yang, Jun Li
School of Computer Science, Sichuan University, Chengdu 610065, China; College of Information Engineering, Sichuan Agricultural University, Yaan 625014, China; School of Aeronautics and Astronautics, Sichuan University, Chengdu 610065, China
 全文: PDF 
摘要： 目的：利用深度卷积网络的优势，解决头部姿态估计中各种关键难点，并提高分类正确率。
创新点：将人工智能的新兴方法深度卷积网络应用在头部姿态估计问题上，根据姿态估计的具体问题设计一套裁剪人脸的方法，改进卷积网络模型、优化参数，并取得了大幅度的效果提升。
方法：首先，因为深度卷积网络算法对图像旋转、尺度、光照等的鲁棒性，图像预处理阶段仅对图像做简单裁剪（图3），并对比了各种裁剪法对分类正确率的影响（表1）。然后，在训练阶段使用适合姿态估计的数据处理策略，通过少量偏移裁剪框和轻微变化图像尺度来获得更多的训练数据以提升效果，在三种公开数据库上报告了实验结果并与目前取得最好效果的三种方法做了对比（表4）。最后，设计两种不同深度的网络，对比网络深度对效果的影响（表2）。
结论：针对头部姿态估计问题，提出了切实有效的新解决方案，并取得了明显改善的效果。
关键词： 头部姿态估计;  卷积神经网络;  多分类    
Abstract: Head pose estimation has been considered an important and challenging task in computer vision. In this paper we propose a novel method to estimate head pose based on a deep convolutional neural network (DCNN) for 2D face images. We design an effective and simple method to roughly crop the face from the input image, maintaining the individual-relative facial features ratio. The method can be used in various poses. Then two convolutional neural networks are set up to train the head pose classifier and then compared with each other. The simpler one has six layers. It performs well on seven yaw poses but is somewhat unsatisfactory when mixed in two pitch poses. The other has eight layers and more pixels in input layers. It has better performance on more poses and more training samples. Before training the network, two reasonable strategies including shift and zoom are executed to prepare training samples. Finally, feature extraction filters are optimized together with the weight of the classification component through training, to minimize the classification error. Our method has been evaluated on the CAS-PEAL-R1, CMU PIE, and CUBIC FacePix databases. It has better performance than state-of-the-art methods for head pose estimation.
Key words: Head pose estimation    Deep convolutional neural network    Multiclass classification
收稿日期: 2015-04-20 出版日期: 2015-11-04
CLC:  TP391  
服务　　
把本文推荐给朋友
加入引用管理器
E-mail Alert
RSS
作者相关文章　　
Ying Cai
Meng-long Yang
Jun Li

引用本文:

Ying Cai, Meng-long Yang, Jun Li. Multiclass classification based on a deep convolutional network for head pose estimation. Front. Inform. Technol. Electron. Eng., 2015, 16(11): 930-939.

链接本文:

http://www.zjujournals.com/xueshu/fitee/CN/10.1631/FITEE.1500125    或    http://www.zjujournals.com/xueshu/fitee/CN/Y2015/V16/I11/930

[1] Guang-hui Song, Xiao-gang Jin, Gen-lang Chen, Yan Nie. 基于两级层次特征学习的图像分类方法[J]. Front. Inform. Technol. Electron. Eng., 2016, 17(9): 897-906.
[2] Gao-li Sang, Hu Chen, Ge Huang, Qi-jun Zhao. 基于稠密多变量标签的“连续”头部姿态估计方法[J]. Front. Inform. Technol. Electron. Eng., 2016, 17(6): 516-526.