Agrupamento de Pixels para o Reconhecimento de Faces

TL;DR

This paper introduces the Pixel Clustering methodology for face recognition, which groups similar pixels into regions to extract features, demonstrating high accuracy with a limited number of features even on shrunk images.

Contribution

The paper proposes a novel Pixel Clustering technique that effectively enhances face recognition accuracy with fewer features and maintains robustness with limited training data.

Findings

512 features are sufficient for high accuracy recognition

Method remains effective with few training classes

Robust performance on multiple face databases

Abstract

This research starts with the observation that face recognition can suffer a low impact from significant image shrinkage. To explain this fact, we proposed the Pixel Clustering methodology. It defines regions in the image in which its pixels are very similar to each other. We extract features from each region. We used three face databases in the experiments. We noticed that 512 is the maximum number of features needed for high accuracy image recognition. The proposed method is also robust, even if only it uses a few classes from the training set.