Emotion detector capable of identifying 7 of the most important human emotions: angriness, disgust, fear, happiness, neutral, sadness, surprise. This application is based on the 68 facial landmarks computed by the shape predictor from "dlib" Python module. The model for this deep learning application is built using Tensorflow Keras and was trained on the FER-2013 dataset (https://www.kaggle.com/deadskull7/fer2013). The dataset is parsed using Pandas module of Python.
Dataset Type | Accuracy |
---|---|
Training | 73.99% |
Test | 54.22% |