Access Control of Object Detection Models Using Encrypted Feature Maps

TL;DR

This paper demonstrates for the first time that encrypted feature maps can effectively control access to object detection models, extending prior work from classification and segmentation to detection tasks.

Contribution

It introduces a novel application of encrypted feature maps for access control specifically tailored to object detection models.

Findings

Encrypted feature maps effectively restrict unauthorized access.

The approach extends access control methods to object detection.

First demonstration of encrypted feature maps in object detection models.

Abstract

In this paper, we propose an access control method for object detection models. The use of encrypted images or encrypted feature maps has been demonstrated to be effective in access control of models from unauthorized access. However, the effectiveness of the approach has been confirmed in only image classification models and semantic segmentation models, but not in object detection models. In this paper, the use of encrypted feature maps is shown to be effective in access control of object detection models for the first time.