How to Detect Lies with a Machine and Microexpressions

Emotions Deeply Hidden

Micro-expressions on the left and their corresponding macro expressions on the right (

AI Microexpression Recognition

Data Collection

A scene representing what it probably felt like for the subjects of this dataset
An example of a participant from the CASME II dataset, you can see the slight difference in her eyebrows which represents a micro-expression. You can also see just how short these are, as this one lasted only 245 milliseconds.

Feature Extraction

Diagram representing how LBP works (
xy, xt, and yt planes

Feature Identification

MicroExpFuseNet vs MicroExpSTCNN

Diagram of how the MicroExpSTCNN model was structured
Diagram of how the MicroExpFuseNet model is structured

Support Vector Machine vs Extreme Learning Machine

SVM tries to find the line that maximizes the distance between the red line and the dotted lines on both sides while accurately dividing the data points into two classes.
SVM in a three-dimensional graph, using a hyperplane to separate the data.


  1. Extreme Learning Machine (97.65%)

Applications of a Micro-Expression Detector

Mental Health

Representation of what people with Smiling Depression do, they mask their feelings with a smile


Quick Recap

New Challenge


  • Microexpressions are short and subtle expressions on the face that usually form in high-stress situations where a person is trying to lie. Good for showing a person's true emotions
  • Using AI we can train models to detect and classify microexpressions
  • A popular dataset right now is the CASME II dataset with 247 data points. Current datasets are sparse and were made under controlled environments.
  • To extract the microexpressions from a face, the machine uses Local Binary Pattern models which compare the grayscale value of the surrounding pixels with its center pixel. This process helps determine edges in a photo. LBP-TOP is used to incorporate time into the model which allows for the model to see changes in the face, more specifically, microexpressions.
  • Two studies created models of their own. The MicroExpSTCNN, which trained off the whole face, and the MicroExpFuseNet which trained off the eye and mouth regions were tested in one study while a Support Vector Machine algorithm and an Extreme Learning Machine algorithm was tested in the other.
  • The ELM was shown to have the highest accuracy of the four at 97.65%. It likely did better than the MicroExpSTCNN and MixroExpFuseNet due to its inclusion of an “Other” category which was used for the more ambiguous microexpressions.
  • Microexpression detection could be used in fields such as Law and Psychiatry where being able to detect deception is crucial.




I am a Undergraduate and TKS innovator at Las Vegas. I am interested in Nanotechnology, Philosophy and Physics.

Love podcasts or audiobooks? Learn on the go with our new app.

Recommended from Medium

Imaginary Soundscape — Take a walk in soundscapes “imagined” by AI

Responsible And Ethical AI — Building Explainable Models

Artificial Intelligence: Family Pet or Future Owners?

Semiconductor Companies Have Been Key Enablers of the AI Revolution.

AI Governance in South East Asian Cities and Governments

Aigents + Reddit — The Next Step Towards “Heternet”

Scenario of RPA in Banking

RPA in Banking

Veterinary Artificial Insemination Market Analysis, Competitive Insight And Key Drivers; Research…

Get the Medium app

A button that says 'Download on the App Store', and if clicked it will lead you to the iOS App store
A button that says 'Get it on, Google Play', and if clicked it will lead you to the Google Play store
Strad Slater

Strad Slater

I am a Undergraduate and TKS innovator at Las Vegas. I am interested in Nanotechnology, Philosophy and Physics.

More from Medium

ML use cases in E-LEARNING

Well, this is me, folks!

Start Your First Machine Learning Project with the Iris flower classification challenge

Image source

How to Adjust Image Contrast using Python Pillow — Change contrast of the image in PIL — python