English (United Kingdom)

https://curated-unify.zendy.io/wp-json/zendy-region/v1/featured_content/oa?rat=en

https://curated-unify.zendy.io/wp-json/zendy-region/v1/highlighted_journal/

Global: Zendy Tools annual

Presents the keyphrase highlighting as premium feature

Keyphrase Highlighting

Presents the summarisation as premium feature

Summarisation

Insights

Presents the pdf analysis as premium feature

PDF Analysis

Presents the zaia usage as premium feature

ZAIA

Global: Zendy Tools monthly

Global: Zendy Plus monthly

Presents the access of premium content as premium feature

Premium Content

Global: Zendy Plus annual

Global: Zendy Free Plan

Stereo matching neural networks often involve a Siamese structure to extractintermediate features from left and right images. The similarity between theseintermediate left-right features significantly impacts the accuracy ofdisparity estimation. In this paper, we introduce a novel adversarial attackapproach that generates perturbation noise specifically designed to maximizethe discrepancy between left and right image features. Extensive experimentsdemonstrate the superior capability of our method to induce larger predictionerrors in stereo neural networks, e.g. outperforming existing state-of-the-artattack methods by 219% MAE on the KITTI dataset and 85% MAE on the Scene Flowdataset. Additionally, we extend our approach to include a proxy networkblack-box attack method, eliminating the need for access to stereo neuralnetwork. This method leverages an arbitrary network from a different visiontask as a proxy to generate adversarial noise, effectively causing the stereonetwork to produce erroneous predictions. Our findings highlight a notablesensitivity of stereo networks to discrepancies in shallow layer features,offering valuable insights that could guide future research in enhancing therobustness of stereo vision systems.

Left-right Discrepancy for Adversarial Attack on Stereo Networks