Evaluation of SVM and kNN classifiers using different data representation methods on Labeled Faces in the Wild dataset
This project aims to evaluate the performance of Support Vector Machines (SVM) and k-Nearest Neighbors (kNN) classifiers using different data representation methods on the Labeled Faces in the Wild (LFW) dataset.
Clone or download this repository.
git clone git@github.com:SurajTC/lfw-face-recognition.git-
Open the
Jupyter NotebookfileLFW_classification.ipynbin Google Colab. -
Run each cell in the notebook to reproduce the results.
Note: Iternet connection is required to download and run the project.
The study demonstrates that the choice of data representation method can significantly affect the performance of the classifiers. The results suggest that Factor Analysis is a promising data representation method for achieving high accuracy and F1-scores with both SVM and kNN classifiers.
This project provides insights into the performance of SVM and kNN classifiers using different data representation methods on the LFW dataset. It highlights the importance of choosing an appropriate data representation method when training classifiers for image recognition tasks.