Showing papers on "Residual frame published in 2019"

Video tamper detection based on multi-scale mutual information

Abstract: Video frame manipulation has become commonplace with the growing easy access to powerful computing abilities. One of the most common types of video frame tampers is the copy-paste tamper, wherein a region from a video frame is replaced with another region from the same frame. In order to improve the robustness of passive video tampering detection, we propose a content-based video similarity tamper passive blind detection algorithm based on multi-scale normalized mutual information which can implement video frame copy, frame insertion and frame deletion tamper detection. The detail implementation of the proposed algorithm consists of multi-scale content analysis, single-scale content similarity measure, multi-scale content similarity measure, and tampering positioning. Firstly, we get the scales of the visual content of the video frame using Gaussian pyramid transform; Secondly, to measure the similarity of single-scale visual content, we define adjacent normalized mutual information of two frames according to information theory; Thirdly, we construct the multi-scale normalized mutual information descriptors to achieve the multi-scale visual content similarity measure of adjacent two frames using a linear combination. Finally, we use the local outlier isolated factor detection algorithm to detect the position of the video tampering. Experimental results show that the proposed approach can not only detect the video frame tampering position of delete, copy, and insert effectively, but also can detect the tampering of different and homology video encoding formats. We obtain a feature detecting accuracy in excess of 93% and detection rate of 96% across post processing operations, and are able to detect the delete, copy, and insert regions with a high true positive rate and lower false positive rate than the existing time field tamper detection methods.

Efficient reference frame compression scheme for video coding systems: algorithm and VLSI design

Abstract: One of the most concerning issues in current video coding systems relies on the bottleneck caused by the intense external memory access required by motion estimation. As memory access affects directly the energy consumption, this problem becomes more evident when battery-powered devices are considered. In this sense, this article presents the Double Differential Reference Frame Compressor (DDRFC), which is a low-complexity and lossless solution to compress the reference frames before storing them in the external memory. The DDRFC performs intra-block double differential coding over 64 × 64-sample blocks to prepare the data for a static Huffman coding. The DDRFC guarantees block-level random access by avoiding data dependencies between neighboring blocks. It reaches an average compression ratio of 69 % for luminance samples for 1080 p video sequences, outperforming any lossless reference frame compressor available in the current literature. Hardware architectures for both the DDRFC encoder and decoder were designed and synthesized targeting FPGA and ASIC 180 and 65-nm standard cells libraries. The synthesis results show that with 65 nm, the DDRFC architectures are able to process 2160 p video at 30 FPS or 1080 p at 120 FPS with a power dissipation of 5.01 mW. The DDRFC codec reaches more than 68 % of energy savings when considering memory communication for HD and UHD video processing.

High-embedding-capacity video steganography method and system based on time sequence residual convolution modeling

Abstract: The invention relates to a high-embedding-capacity video steganography method and system based on time sequence residual convolution modeling. The method comprises the following steps: marking a reference frame and a residual frame of a secret video, simultaneously processing the reference frame and the residual frame by adopting a Y-shaped convolutional neural network to hide secret information and output a carrier video frame, and synthesizing the carrier video frame into a carrier video; and recovering secret information in the carrier video by adopting a Y-type convolutional neural network. Compared with a convolutional neural network-based image steganography algorithm which is directly applied to video steganography, the video steganography method has the advantages that the sparsityof a residual error between continuous frames is explored, two Y-type convolutional neural network structures are adopted, different endpoint processing is adopted for video frames with different properties, and part of convolutional layer parameters are shared at the same time. One video can be hidden in the other video with the same length, the hidden information amount can reach 24 bpp and ismuch larger than that of a traditional method, and the problem that the traditional method cannot be applied to high-embedding-capacity video steganography is solved to a great extent.

Residual Frame for Noisy Video Classification According to Perceptual Quality in Convolutional Neural Networks

Abstract: Perceptual quality of a video describes the quality consistent with human perception. The growing popularity of short video sharing on mobile platforms such as Tik Tok and WeSee makes the video assessment system based on perceptual quality a necessity. In practice, short videos captured by mobile devices often contain different types of distortions incurred by sensor noise or compression noise, which potentially makes the videos visually unpleasing to users and may degrade the performance of deep neural networks when applied to these noisy videos. Thus, it is necessary to identify noisy videos based on video perceptual quality. However, traditional video/image noise estimation methods are designed to estimate the variance of homogeneously distributed synthetic noise, not real noise. In this paper, we propose a simple yet effective method to recognize the noisy videos using their residual frames. Since the original video frame contains rich content information, which may result in under-or over-estimation of the noise, we construct residual frames to reduce the influence of the content information while maintaining the main noise information in the video. We also create a new data set with more than 30 thousand images captured from videos with real noise. Experimental results demonstrate the effectiveness of our proposed method.

Video encoding apparatus and operating method thereof

Abstract: A video encoding apparatus and an operating method thereof are provided. The video encoding apparatus includes an encoding circuit, a reconstructed frame generating circuit, a deblocking filter and a determination circuit. The encoding circuit generates encoded data according to a residual frame and generates a reconstructed residual frame. The reconstructed frame generating circuit generates a first reconstructed frame according to the reconstructed residual frame and a predicted frame. The deblocking filter generates a second reconstructed frame by eliminating discontinuities in reconstructed blocks of the first reconstructed frame. The current reconstructed block is not output to the memory when a current reconstructed block is identical to a co-located reference block of a reference frame stored in the memory. The current reconstructed block is output to the memory when the current reconstructed block is different from the co-located reference block in the reference frame.

Parametric Approach of Modeling the Source Signal

Abstract: In this chapter, two parametric source modeling methods are proposed for improving the quality of HMM-based speech synthesis. The two methods model the pitch-synchronous residual frames extracted from the excitation signal based on principal component analysis. In the first method, the pitch-synchronous residual frames are parameterized using principal component analysis. Every residual frame is represented using 30 PCA coefficients. In the second method, an analysis of characteristics of the residual frames around GCI is performed using PCA. Based on the analysis, the pitch-synchronous residual frames are decomposed into deterministic and noise components.