Visual Decoding of Hidden Watermark in Trained Deep Neural Network | IEEE Conference Publication | IEEE Xplore

Visual Decoding of Hidden Watermark in Trained Deep Neural Network


Abstract:

Deep neural network (DNN) is a kind of intellectual property considering its usefulness and cost to develop. This paper proposes watermarking to a trained DNN models to p...Show More

Abstract:

Deep neural network (DNN) is a kind of intellectual property considering its usefulness and cost to develop. This paper proposes watermarking to a trained DNN models to protect its copyright. The proposed method has a remarkable feature for watermark detection process, which can decode the embedded pattern cumulatively and visually. In the experiment, we can embed a specific visual pattern using 5,000 or 60,000 images on the pretrained image classification DNN model. Then the embedded pattern is decoded using 20 images out of the 5,000 or 60,000 images, while a performance degradation to the original image classification task is small. At the conference site, real time and animated visual decoding demonstration is performed.
Date of Conference: 28-30 March 2019
Date Added to IEEE Xplore: 25 April 2019
ISBN Information:
Conference Location: San Jose, CA, USA

I. Introduction

DNN (Deep Neural Network) has attained much interests from both researchers and application developers because of its high performance for classification, regression, and prediction. To achieve such performance, it is necessary to use huge data sets and computing for many hours. Considering it costs much money and time, it should be regarded as a kind of intellectual property [1].

References

References is not available for this document.