This project implements a deep learning model for the detection of human emotions in images. It uses Convolutional Neural Networks (CNNs) to classify images into three emotional categories: angry, happy, and sad.

The dataset used in this project is the "Human Emotions Dataset" available on Kaggle. It consists of labeled images representing different emotions.

Human Emotions Dataset

The model architecture is based on a modified LeNet-5 CNN architecture. It includes layers for data preprocessing, convolution, batch normalization, max-pooling, dropout, and fully connected layers. Regularization techniques are used to prevent overfitting.

The project is configured with various hyperparameters, including batch size, image size, learning rate, and dropout rate, which can be adjusted as needed.

The model is trained using TensorFlow and Keras. The training dataset is shuffled and preprocessed, and the model's performance is evaluated using a validation dataset.

To use this project, follow these steps:

1. Clone the repository:

2. Install the required libraries:

pip install -r requirements.txt

3. Download and unzip the dataset (or use your own dataset) into the appropriate directories:

• Training data: /content/dataset/Emotions Dataset/Emotions Dataset/train
• Validation data: /content/dataset/Emotions Dataset/Emotions Dataset/test

4. Adjust the configuration in the code as needed.

5. Train the model.

6. Evaluate the model.

7. Make predictions on new data:

The model achieved a significant level of accuracy in classifying human emotions in images. Training and validation metrics, such as accuracy and loss, can be found in the training logs.

Successful implementation of a modified LeNet-5 architecture for emotion detection. Efficient preprocessing of image data, including resizing and rescaling. Effective use of regularization techniques to prevent overfitting. Deployment of a TensorFlow-based deep learning model.

Mention any challenges you encountered during the project, such as data quality issues, overfitting, or performance optimization.

Provide insights into potential future improvements or enhancements for the project. Ideas may include fine-tuning hyperparameters, exploring different architectures, or incorporating real-time emotion detection from a video stream.

Express gratitude to any contributors, libraries, or datasets that played a significant role in achieving the project's outcomes.

Include any relevant visualizations, such as confusion matrices or ROC curves, to showcase the model's performance.

• (https://github.com/biswadeep_roy)

This project is licensed under the MIT License - see the MIT file for details.

Feel free to reach out if you have any questions or suggestions!