Countering Inconsistent Labelling by Google’s Vision API for Rotated Images

Document Type

Conference Proceeding

Publication Title

Advances in Intelligent Systems and Computing

Abstract

Google’s Vision API analyses images and provides a variety of output predictions, one such type is context-based labelling. In this paper, it is shown that adversarial examples that cause incorrect label prediction and spoofing can be generated by rotating the images. Due to the black-boxed nature of the API, a modular context-based pre-processing pipeline is proposed consisting of a ResNet50 model that predicts the angle by which the image must be rotated to correct its orientation. The pipeline successfully performs the correction whilst maintaining the image’s resolution and feeds it to the API which generates labels similar to the original correctly oriented image, and using a percentage error metric, the performance of the corrected images as compared to its rotated counterparts is found to be significantly higher. These observations imply that the API can benefit from such a pre-processing pipeline to increase robustness to rotational perturbances.

First Page

202

Last Page

213

DOI

10.1007/978-981-15-6067-5_23

Publication Date

1-1-2021

This document is currently not available here.

Share

COinS