Gadak

Full-stack facial-attendance system with real-time identity verification — basis for a published study on facial-recognition distance metrics.

Gadak is a full-stack facial-attendance system built with FastAPI and Angular, supporting reliable real-time identity verification.

Stack: Python, FastAPI, PostgreSQL, Docker, Scikit-learn, Jupyter.

Key work:

Developed facial-recognition modules using HOG features, Linear SVM, and KNN classifiers.
Trained and tuned models with Scikit-learn; evaluated with Matplotlib and Jupyter notebooks.
The evaluation of model trade-offs under dataset constraints led to a peer-reviewed publication in the Journal of Science and Technology (Chhetri & Kshetri, 2024).

References

2024

JoST
Decoding Facial Recognition: Analyzing Standalone Euclidean and KNN Distance Metrics

Prajwol Chhetri and Sunil Raut Kshetri

Journal of Science and Technology, 2024

Abs DOI Bib HTML PDF

This study compares the performance of K-Nearest Neighbors (KNN) using different distance metrics—Euclidean, Manhattan, and Minkowski—as well as the standalone Euclidean distance metric in facial recognition tasks, evaluating accuracy and computational efficiency across diverse datasets including Celebrity Faces, Color FERET, Family Faces, and Yale Faces.
@article{chhetri2024decoding, title = {Decoding Facial Recognition: Analyzing Standalone Euclidean and KNN Distance Metrics}, author = {Chhetri, Prajwol and Kshetri, Sunil Raut}, journal = {Journal of Science and Technology}, volume = {4}, number = {2}, pages = {51--57}, year = {2024}, doi = {10.3126/jost.v4i2.78952}, }