Mood detection
It identifies your mood level. Negative, positive, or zero.
Description:
- Uses opencv library to detect and crop the face.
- Uses a modified MobileNetV2 for mood regression. (mood level is any number between -1 and 1)
Model
Model training is in the notebook.
- The model is initially trained on FER dataset.
- Then modifying the head, it is trained (transfer learning) to my mood regression personal dataset.