简体   繁体   English

情绪识别或特征提取

[英]Emotion Recognition or Feature Extraction

I am working on a module that can recognize emotion from human faces (basic emotion such as : joy, sad, disgust, fear, surprise, neutral). 我正在开发一个可以识别人脸情绪的模块(基本的情绪,例如:欢乐,悲伤,厌恶,恐惧,惊奇,中立)。

The tools I am using include : 我使用的工具包括:

  1. OpenCV 2.4.2 OpenCV 2.4.2
  2. Qt Creator Qt创作者
  3. Mingw 4.4 Mingw 4.4

First I thought I can make use of Open CV s Face Recognition module to suit my purpose and played a with it (LBPH Face recognizer), but could achieve much results. 首先,我认为我可以利用Open CV的人脸识别模块来适应我的目的并使用它(LBPH人脸识别器),但是可以取得很多效果。 I trained the recognizer with images from different classes of emotions and do a face recognition for a sample set of image to recognize the emotion (not the person). 我用来自不同类别情感的图像训练了识别器,并针对图像样本集进行了人脸识别以识别情感(而不是人)。

Then I dropped that idea and started searching for feature extraction methods so that once I get a sample facial image I can classify it based on the extracted features (using SVM). 然后我放弃了这个想法,开始寻找特征提取方法,以便一旦获得样本面部图像,就可以基于提取的特征(使用SVM)对其进行分类。 I came across some libraries but all are meant for Visual Studio. 我遇到了一些库,但所有库都用于Visual Studio。

Is that I should change the whole project to work with visual studio?...Is there any other library which can help me to extract facial feature (with the current set of tools I am working with)...Or Am I missing somethin?? 是否应该更改整个项目以使其与Visual Studio一起使用?...是否有其他库可以帮助我提取面部特征(使用当前使用的工具集)...或者我缺少某些功能? ??

Thanks 谢谢

Can make use of Dlib and SVM for the purpose. 可以利用Dlib和SVM。 Get the pose of face using DLib shape prediction module and pass the the feature points(distance between upper lip,lower lip, mouth width etc) to SVM to predict the emotion. 使用DLib形状预测模块获取脸部姿势,并将特征点(上嘴唇,下嘴唇之间的距离,嘴巴宽度等)传递到SVM以预测情绪。 Note that the SVM predictor should be trained using sample images and their feature points 请注意,应使用样本图像及其特征点对SVM预测器进行训练

声明:本站的技术帖子网页,遵循CC BY-SA 4.0协议,如果您需要转载,请注明本站网址或者原文地址。任何问题请咨询:yoyou2525@163.com.

 
粤ICP备18138465号  © 2020-2024 STACKOOM.COM