Hak-Yeol Choi

Stereoscopic watermarking by horizontal noise mean shifting

Depth-image-based rendering (DIBR) is a method to represent a stereoscopic content. The DIBR consists of a monoscopic center view and an associated per-pixel depth map. Using these two components and given depth condition from a user, the DIBR renders left and right views. The advantages of DIBR are numerous. The user can choose not only the monoscopic or stereoscopic view selectively, but also the depth condition what he prefers when he watches a stereoscopic content. However, in the view of copyright protection, since not only the center view but also each left or right view can be used as a monoscopic content when they are illegally distributed, the watermark signal which is embedded in the center view must have an ability to protect the respective three views. In this study, we solve this problem by exploiting the horizontal noise mean shifting (HNMS) technique. We exploit the fact that the objects in the view are shifted only to horizontal way when the center view renders to the left and right views. Using this fact, the proposed stereoscopic watermarking scheme moves the mean of horizontal noise histogram which is invariant to horizontal shifting, and we achieve good performance as shown in the experimental results.

Cropping-resilient 3D mesh watermarking based on consistent segmentation and mesh steganalysis

This paper presents a new approach to 3D mesh watermarking using consistent segmentation and mesh steganalysis. The method is blind, statistical, and highly robust to cropping attack. The primary watermarking domain is calculated by shape diameter function and the outliers of segments are eliminated by computing the consistency interval of vertex norms. In the watermark embedding process, the mesh is divided into several segments and the same watermark is inserted in each segment. In the watermark extraction process, the final watermark among watermark candidates extracted from multiple segments is determined through watermark trace analysis that is kind of mesh steganalysis. We analyze the watermark trace energy of multiple segments of a mesh and detect the final watermark in the segment with the highest watermark trace energy. To analyze the watermark trace energy, we employ nonlinear least-squares fitting. The experimental results show that the proposed method not only achieves significantly high robustness against cropping attack, but also resists common signal processing attacks such as additive noise, quantization, smoothing and simplification.

Detecting composite image manipulation based on deep neural networks

In this paper, we propose a composite manipulation detection method based on convolutional neural networks (CNNs). To our best knowledge, this is the first work applying deep learning for composite forgery detection. The proposed technique defines three types of attacks that occurred frequently during image forging and detects when they are concurrently applied to images. To do this, we learn the statistical change due to the manipulation through the proposed CNN architecture and classify the manipulated image. The proposed technique is effective since it learns integrated image of composite manipulation and extracts characteristic distinguished from original image. Since most attacks are applied in a composite way in real environment, the approach of the proposed technique has practical advantages compared to traditional forensics scheme. In addition, the experimental results demonstrate the reliability of the proposed method through results of high performance.

Improved 3D Mesh Steganalysis Using Homogeneous Kernel Map

Steganalysis targets to detect the existence of hidden information in a given content. In this paper we propose to use a local feature set which is designed to enhance discrimination of features obtained from a cover and a stego mesh. The proposed feature captures the fine deformation of the 3D mesh surface induced by a steganography or watermarking method. In our 3D steganalysis approach, in addition, we apply the homogeneous kernel map to the local feature set, which make it possible to bring much more discrimination via non-linear mapping. The proposed feature set and its combination with the homogeneous feature map have shown good performance on two different steganography and watermarking algorithm with a well known and widely used 3D mesh database through repeated experiments.

Fingerprint Spoof Detection Using Contrast Enhancement and Convolutional Neural Networks

Recently, as biometric technology grows rapidly, the importance of fingerprint spoof detection technique is emerging. In this paper, we propose a technique to detect forged fingerprints using contrast enhancement and Convolutional Neural Networks (CNNs). The proposed method detects the fingerprint spoof by performing contrast enhancement to improve the recognition rate of the fingerprint image, judging whether the sub-block of fingerprint image is falsified through CNNs composed of 6 weight layers and totalizing the result. Our fingerprint spoof detector has a high accuracy of 99.8% on average and has high accuracy even after experimenting with one detector in all datasets.

Perceptual 3D Watermarking Using Mesh Saliency

In this paper, we introduce a novel blind 3D mesh watermarking method which focuses on preserving the appearance of the watermarked model. Despite the high transparency achieved by existing 3D watermarking schemes, we observe that only a small amount of geometric error can bring a significant impact to appearance of 3D models, especially in visually important regions. We integrate this human perceptual importance, called saliency, to control the distortions on surfaces. Our method enhances the imperceptibility while maintaining the efficiency of processing spatial information by conjugating spatial and spectral regions. We use the vertex norm distribution and solve the quadratic error minimization problem to insert watermark bits. Experimental results demonstrate that our method performs well for perceived visual quality and also robustness against various geometric attacks.

Content Recapture Detection Based on Convolutional Neural Networks

Detecting recaptured images has been considered as an important issue. The previous techniques tried to make hand-crafted features represent the statistical characteristics of the recaptured images. Different to the existing methods, the proposed method solves the recapturing detection problem based on a deep learning technique which shows high performance for various applications in recent image processing. Specifically, we propose a recaptured image classification scheme based on a convolutional neural networks (CNNs). To our best knowledge, this is the first work of applying CNNs into the recaptured image detection. For reliable performance evaltuation, we used high-quality database for training and testing. The experimental results show high performance compared to the state-of-the-art methods.

Blind 3D mesh watermarking based on cropping-resilient synchronization

This paper proposes a novel anti-cropping blind 3D mesh watermarking method. Although there have been many mesh watermarking studies, methods with robustness to cropping attack are rare. Existing anti-cropping watermarking methods show only slight robustness to cropping and signal processing attack. Cropping is one of the most severe attack, since it significantly undermines the synchronization of watermarking. In this paper, we solve the synchronization problem in a blind environment which is a core part of the anti-cropping watermarking method using local shape-based synchronization. The proposed local shape-based synchronization is robust to not only cropping attack, but also similarity and distortion attack, since it uses the shape information of the mesh, not the surface information. In the watermark embedding process, the distortion from the watermark was minimized by using the method of spreading based on segmented bin. Additionally, the method has higher security than the existing method. In the experimental results, the proposed method shows high robustness against common signal processing attacks as well as severe cropping attacks with high invisibility.

Blind 3D Mesh Watermarking Based on Sphere-Shape Template

In this paper, we propose a novel blind 3D mesh watermarking scheme based on synchronization using a template. The proposed template has a sphere shape with a basis point and scale information for the watermarking. The template is invariant to translation, rotation, and scale. In the extraction process of the template, the post-processing, which sorts out the genuine template vertices from many candidate vertices, greatly increases the extracting accuracy. The proposed watermarking method is designed to extract the watermark from the rest of the region even if part of the mesh information is cropped. According to the experimental results, the proposed method has a higher robustness against not only cropping but also general signal processing attacks than the previous methods.

DIBR 3D VIDEO WATERMARKING WITH FASTER DT-CWT QUANTIZATION

As the 3D content market is growing up rapidly, copyright issues on the 3D content naturally have come into play. Among the various 3D coding techniques, depth-imagebased rendering (DIBR) has been studied by numerous researchers due to its coding efficiency and 2D compatibility. However, conventional watermarking scheme for 2D content cannot be directly applied to the DIBR content because DIBR process is considered as a new type of attack in terms of watermarking design. In this paper, we propose a watermarking scheme for DIBR 3D videos with an improved version of DT-CWT magnitude quantization. In order for efficient watermark embedding and extraction on DIBR 3D videos, all frames are grouped. Then, the watermark is embedded repeatedly and periodically so that extraction performance is enhanced. Furthermore, the quantization process is improved to a faster version by checking sum of magnitude and setting a start point of quantization order optimally. The simulation results show that the embedded watermark is stably extracted from center views and virtual views. Moreover, when the watermarked virtual views are distorted or processed by compression, scaling, median filtering and noise addition, the embedded watermark is extracted with low BERs.