Abstract:In recent years, many model intellectual property (IP) proof methods for IP protection have been proposed, such as model watermarking and model fingerprinting. However, with the increasing number of models transmitted and deployed on the Internet, quickly finding the suspect model among thousands of models on model-sharing platforms such as GitHub is in great demand, which concurrently triggers the new security problem of model copy detection for IP protection. As an important part of the model IP protection system, the model copy detection task has not received enough attention. Due to the high computational complexity, both model watermarking and model fingerprinting lack the capability to efficiently find suspected infringing models among tens of millions of models. In this article, inspired by the hash-based image retrieval methods, we introduce a novel model copy detection mechanism: perceptual hashing for convolutional neural networks (CNNs). The proposed perceptual hashing algorithm can convert the weights of CNNs to fixed-length binary hash codes so that the lightly modified version has the similar hash code as the original model. By comparing the similarity of a pair of hash codes between a query model and a test model in the model library, similar versions of a query model can be retrieved efficiently. To the best of our knowledge, this is the first perceptual hashing algorithm for deep neural network models. Specifically, we first select the important model weights based on the model compression theory, then calculate the normal test statistics (NTS) on the segments of important weights, and finally encode the NTS features into hash codes. The experiment performed on a model library containing 3,565 models indicates that our perceptual hashing scheme has a superior copy detection performance.

Efficient Hashing Method Using 2D-2D PCA for Image Copy Detection

Secure and robust image perceptual hashing

Robust Hashing via Global and Local Invariant Features for Image Copy Detection

Efficient Image Hashing with Geometric Invariant Vector Distance for Copy Detection

Robust and fast image hashing with two-dimensional PCA

Weakly Principal Component Hashing with Multiple Tables.

A Novel Hashing Scheme Via Image Feature Map and 2D PCA

Robust Hashing with Deep Features and Meixner Moments for Image Copy Detection

Perceptual Image Hashing with Locality Preserving Projection for Copy Detection

Robust Hashing With Local Tangent Space Alignment for Image Copy Detection

Robust Image Hashing With Isomap and Saliency Map for Copy Detection

Geometrically Robust Video Hashing Based on ST-PCT for Video Copy Detection

Feature aggregating hashing for image copy detection

Fine-search for image copy detection based on local affine-invariant descriptor and spatial dependent matching

A Novel Block-DCT and PCA Based Image Perceptual Hashing Algorithm

Detection Algorithm of Image Forgery Based on Principal Components Analysis of Projection Data

CNN Feature-Based Image Copy Detection with Contextual Hash Embedding

Robust video hashing based on deep feature and quaternion generic Fourier descriptor for copy detection

Robust Image Hashing for Image Authentication Based on DCT-DWT Composite Domain

A Dct-Vq Based Multipurpose Image Hashing Scheme For Copyright Protection And Content Authentication

Perceptual Hashing of Deep Convolutional Neural Networks for Model Copy Detection