Attacking Optical Character Recognition (OCR) Systems with Adversarial Watermarks
This addresses the vulnerability of OCR systems to adversarial attacks in a way that evades human detection, though it is incremental as it builds on existing attack methods.
The paper tackles the problem of generating adversarial examples for OCR systems that are unnatural and easily detectable by proposing a watermark attack method to create natural distortions disguised as watermarks, achieving similar attack performance to state-of-the-art methods in various scenarios.
Optical character recognition (OCR) is widely applied in real applications serving as a key preprocessing tool. The adoption of deep neural network (DNN) in OCR results in the vulnerability against adversarial examples which are crafted to mislead the output of the threat model. Different from vanilla colorful images, images of printed text have clear backgrounds usually. However, adversarial examples generated by most of the existing adversarial attacks are unnatural and pollute the background severely. To address this issue, we propose a watermark attack method to produce natural distortion that is in the disguise of watermarks and evade human eyes' detection. Experimental results show that watermark attacks can yield a set of natural adversarial examples attached with watermarks and attain similar attack performance to the state-of-the-art methods in different attack scenarios.