How shayanali005/facial-expression-cnn-resnet50- works
This repository presents a notebook-based facial-expression classifier: grayscale images are normalized, passed through a CNN, and evaluated across seven emotion classes. The README documents preprocessing, the model blueprint, batch training with Adam, and accuracy/loss and confusion-matrix diagnostics. The notebook is the only listed executable artifact, but its source was not provided, so the graph reflects documented capabilities rather than verified call structure.
Main entry points: Facial_expression.ipynb Image preprocessing
Components 7
Image preparation
- Image preprocessing
Facial_expression.ipynb
Model and training
- CNN classifier
Facial_expression.ipynb - Model training
Facial_expression.ipynb
Evaluation
- Performance metrics
Facial_expression.ipynb - Confusion matrix
Facial_expression.ipynb
Other components
- Analyst
- Facial images
Connections 6
- Analyst → Image preprocessing: runs
- Facial images → Image preprocessing: provides images
- Image preprocessing → CNN classifier: feeds normalized images
- CNN classifier → Model training: trains model
- Model training → Performance metrics: evaluates model
- Performance metrics → Confusion matrix: summarizes errors