Human-Emotion-Detection-Through-Computer-Vision

Human Emotion Detection Repository \====================================== This repository is structured to provide a comprehensive overview of our Human Emotion Detection project, including data preparation, model architectures, training details, and performance analysis. Below is a detailed explanation of the repository structure: Repository Structure 1\. data/ Contains all datasets in TFRecord format, organized into three categories: raw/: Original dataset without any augmentations. augmented/training/: Dataset with standard augmentations applied. cut\mix\augmented/training/: Dataset with CutMix augmentation, enhancing data diversity. 2\. models/ Stores training history, architectural details, and metrics for each model used in the project: EfficientNetB4/: Files related to the EfficientNetB4 architecture, including updates and fine-tuning details. EfficientNetB4\finetuned/: Files and performance metrics for the fine-tuned EfficientNetB4 architecture. LeNet-5/: Contains files for the LeNet-5 model and its performance metrics. ResNet-34/: Includes files and results for the ResNet-34 model. 3\. notebooks/