Face Recognition for 31 different classes

[aaradhyasinghgaur]

Is your feature request related to a problem? Please describe. Face recognition for 31 different classes involves identifying or verifying individuals from a set of 31 distinct faces. Challenges include managing intra-class variations such as changes in lighting, facial expressions, and angles, as well as inter-class similarities where different individuals may have similar facial features. Deep learning models can address these issues by learning robust feature representations that discriminate between the 31 classes, ensuring high accuracy even under varied conditions.

Describe the solution you'd like Using various different pre-trained models such as Xception , ResNet50 , VGG16 etc . I want to analyse , research and determine the most suitable model after modifying it using data-augmentation techniques and comparing it using performance matrices such as accurcay score , bar and graph plotting and also doing EDA analysis on the data.

Dataset I'll use :- https://www.kaggle.com/datasets/vasukipatel/face-recognition-dataset

Describe alternatives you've considered A clear and concise description of any alternative solutions or features you've considered.

Approach to be followed (optional) 1.) Using 5 different available network architectures such as DenseNet121 , Xception , VGG16 etc. to face recogniion task. 2.) use data augmentation techniques to improve the accuarcy of models. 3.) Comparing performance and accuracy of models using accuracy score ,loss and accuracy graph , confusion matrix for better understanding. 4.) Perfroming EDA (data analysis) for dataset to understand the structure of data. 5.) Using README file for describing the work I've performed.

Additional context Add any other context or screenshots about the feature request here.

@akshitagupta15june kindly assign this issue to me.

[github-actions[bot]]

Hello @kyra-09, Thank you for generating an issue to this project! Please wait while we get back to you.

[Rukkya]

hello @kyra-09 I'd like to work on the same issue im pretty enthusiastic about the idea , please let me know if possible

[aaradhyasinghgaur]

hey @akshitagupta15june kindly assign this issue to me with the suitable label so that I can start working on it.

[akshitagupta15june]

hey the algorithms are already present in repo