Generation of Gradient-Preserving Images allowing HOG Feature Extraction

TL;DR

This paper introduces a method for creating gradient-preserving images that enable privacy-preserving HOG feature extraction for machine learning applications like face recognition.

Contribution

The paper presents a novel image generation technique that preserves gradients, facilitating privacy-preserving feature extraction for machine learning tasks.

Findings

HOG features from gradient-preserving images are effective for face recognition.

The method enables privacy-preserving feature extraction without significant loss of recognition accuracy.

Experimental results demonstrate the practicality of the approach.

Abstract

In this paper, we propose a method for generating visually protected images, referred to as gradient-preserving images. The protected images allow us to directly extract Histogram-of-Oriented-Gradients (HOG) features for privacy-preserving machine learning. In an experiment, HOG features extracted from gradient-preserving images are applied to a face recognition algorithm to demonstrate the effectiveness of the proposed method.