Human face, emotion and voice recognition

问题:

I am looking for a good face, emotion and voice recognition method in C#. For face recognition I was early using Emgu CV which is not accurate and performance is very low in low light conditions. Also I need to find user's emotion. Whether sad or happy like that. But I found its not easy with Emgu CV.

Also for voice recognition I am not able to find any solutions yet, I found speech recognition but it is not what I need.

I don't want to use any online API's. Can anybody suggest me any SDKs or Algorithms using which I a implement face, emotion and voice recognition?

回答1:

Face recognition

• Here you can fine an introductory pdf, in particular take a look at the references for more details.
• Here you can find a good tutorial and demo program with free source code for performing face detection and face recognition in real-time from a webcam (also in OpenCV).
• Eigenface and Face Recognition Homepage.
• You can download the CSU Face Identification Evaluation System Version 5.1.
• Some articles about face detection.
• From Code Project: Multiple face detection and recognition in real time - Face Detection C++ Library with Skin and Motion Analysis - Face Detection in C# - Face and Eyes Detection Using OpenCV.
  

Voice recognition

Simply take a look at System.Speech.Recognition Namespace (look at this and this answer).

回答2:

It seems that you need Kinect. http://www.microsoft.com/en-us/kinectforwindows/develop/

回答3:

I understand the question is very old, although I would like to contribute to it since I am also working on similar application

For Face, Emotion and Voice recognition there are plenty available now, find below some options

1) Face detection and recognition using OpenCV: http://ahmedopeyemi.com/main/face-detection-and-recognition-in-c-using-emgucv-3-0-opencv-wrapper-part-1/

Pro: Open source so basically free. You can modify the code to accommodate your needs. Multiple platforms supported.

Con: Accuracy is tolerable but not great. This can be resolved by modifying the library to your needs.

2) Face detection and recognition with emotions: try Microsoft Cognition Service API

Pro: very accurate. It includes many other services such as NLP, Speech to Text, Text to Speech, Computer Vision, RT language translator and so on. Accuracy is great since the services are hosted on cloud and are being trained for every data input.

Con: Paid solution. Little expensive. You have to pay for each service separately.

3) Speech recognition: Google Voice API Pro: similar to MS cognition service. Accuracy is better in Noise. con: Paid. May not be good for one time sale commercial product

4) Speech recognition and NLP: wit.ai/api.ai Pro: Free to use. Can use data from other public apps. Accuracy is good. Con: I don't find find any

Try some more links as follows. May help a little.

MS Face API based security system https://www.hackster.io/peejster/facial-recognition-security-system-c6ca6c