Convolutional Neural Networks for Facial Expression Recognition on the FER2013 Dataset.
We aim to appropriately categorise facial expressions into one of the seven categories listed below in this study.
Angry----Disgust----Fear----Happy----Sad----Surprise----Neutral
This URL will take you to the project's dataset: https://www.kaggle.com/deadskull7/fer2013
Download the file, then unzip it.
There are around 32300 photos listed in this single csv file.
The model's greatest accuracy was about 63%.
Emotion_Recognition_Train.ipynb contains the model and the pre-processing procedures.
By pressing this, the webcam will start and the frames collected by our trained model for inference will be fed.
The faces in the frames are identified using a Haarcascade algorithm, and the detected region is then cropped to the required size and provided as input to the detector.
Detector_In_Action.py in Python
By pressing this, the webcam will start and the frames collected by our trained model for inference will be fed.
The faces in the frames are identified using a Haarcascade algorithm, and the detected region is then cropped to the required size and provided as input to the detector.
