ICSSE 2020 : Face Expression and Tone of Voice for Deception System (Award of best student paper)
The overall flowchart for deception detection is illustrated below. We combine many features into a vector and then apply SVM to classification.
Because the face can be many angles, we need to align the face before using it.
Our deception detection system comprises four parts:
- 3D landmarks displacement
- Emotion Unit
- Action Unit
- Emotion Audio unit
Please see the requirements.txt
for more details.
121 videos including 61 deceptive videos and 60 truth videos related paper
325 videos including 162 deceptive videos and 163 truth videos related paper
145 videos including 62 deceptive videos and 83 truth videos
- Note: If you would like access to the above datasets, please contact the authors who provided the dataset, respectively.
python lie_GUI.py
- If you have any implementation problem, feel free to E-mail me! [email protected]