Filters








1,794 Hits in 5.4 sec

Deep Residual Learning in the JPEG Transform Domain [article]

Max Ehrlich, Larry Davis
2019 arXiv   pre-print
We introduce a general method of performing Residual Network inference and learning in the JPEG transform domain that allows the network to consume compressed images as input.  ...  We show that the sparsity of the JPEG format allows for faster processing of images with little to no penalty in the network accuracy.  ...  Conclusion and Future Work In this work we showed how to formulate deep residual learning in the JPEG transform domain, and that it provides a notable performance benefit in terms of processing time per  ... 
arXiv:1812.11690v3 fatcat:mqqwezztrrgmzczqnoq4pmic6y

A Comprehensive Benchmark for Single Image Compression Artifacts Reduction [article]

Jiaying Liu, Dong Liu, Wenhan Yang, Sifeng Xia, Xiaoshuai Zhang, Yuanying Dai
2019 arXiv   pre-print
In this work, a systematic listing of the reviewed methods is presented based on their basic models (handcrafted models and deep networks).  ...  Furthermore, based on a unified deep learning configuration (i.e. same training data, loss function, optimization algorithm, etc.), we evaluate recent deep learning-based methods based on diversified evaluation  ...  Deep learning-based JPEG artifacts removal Deep learning-based methods largely improve the restoration capacity of the data-driven methods.  ... 
arXiv:1909.03647v1 fatcat:yujaixpevzeadi7zfh7ap6t7jq

Investigating Image Applications Based on Spatial-Frequency Transform and Deep Learning Techniques [article]

Qinkai Zheng, Han Qiu, Gerard Memmi, Isabelle Bloch
2020 arXiv   pre-print
This is the report for the PRIM project in Telecom Paris. This report is about applications based on spatial-frequency transform and deep learning techniques.  ...  At the receiver's end, we propose a DC recovery algorithm together with the deep residual learning framework to recover images with high quality.  ...  On one hand, deep learning techniques can be used to enhance the existing method that is based on transformations in the frequency domain, e.g. JPEG compression based on discrete cosine transform.  ... 
arXiv:2004.02756v1 fatcat:6owbiqiwurfxljiaqm7m2zyo6a

DPW-SDNet: Dual Pixel-Wavelet Domain Deep CNNs for Soft Decoding of JPEG-Compressed Images

Honggang Chen, Xiaohai He, Linbo Qing, Shuhua Xiong, Truong Q. Nguyen
2018 2018 IEEE/CVF Conference on Computer Vision and Pattern Recognition Workshops (CVPRW)  
The pixel domain deep network takes the four downsampled versions of the compressed image to form a 4-channel input and outputs a pixel domain prediction, while the wavelet domain deep network uses the  ...  1-level discrete wavelet transformation (DWT) coefficients to form a 4-channel input to produce a DWT domain prediction.  ...  [17] introduced the residual learning technique and designed a very deep network of twenty layers for single image super-resolution.  ... 
doi:10.1109/cvprw.2018.00114 dblp:conf/cvpr/ChenHQXN18 fatcat:rllwqhn2hbdmzckqh5biq7qyxm

DPW-SDNet: Dual Pixel-Wavelet Domain Deep CNNs for Soft Decoding of JPEG-Compressed Images [article]

Honggang Chen and Xiaohai He and Linbo Qing and Shuhua Xiong and Truong Q. Nguyen
2018 arXiv   pre-print
The pixel domain deep network takes the four downsampled versions of the compressed image to form a 4-channel input and outputs a pixel domain prediction, while the wavelet domain deep network uses the  ...  1-level discrete wavelet transformation (DWT) coefficients to form a 4-channel input to produce a DWT domain prediction.  ...  [17] introduced the residual learning technique and designed a very deep network of twenty layers for single image super-resolution.  ... 
arXiv:1805.10558v1 fatcat:sitqnybqmbeexpr36yuoe2yyua

Detection of Plant Leaf Disease Directly in the JPEG Compressed Domain using Transfer Learning Technique [article]

Atul Sharma, Bulla Rajesh, Mohammed Javed
2021 arXiv   pre-print
In this research paper, plant leaf disease detection employing transfer learning is explored in the JPEG compressed domain.  ...  Recently, Deep Neural Networks have been exceptionally fruitful in image classification.  ...  Fig. 6 . 6 The proposed deep architecture for plant leaves disease detection in JPEG compressed domain.  ... 
arXiv:2107.04813v1 fatcat:fvdg33yvsralbom5rttx6sdaz4

Compression Artifacts Removal Using Convolutional Neural Networks [article]

Pavel Svoboda and Michal Hradis and David Barina and Pavel Zemcik
2016 arXiv   pre-print
We were able to train networks with 8 layers in a single step and in relatively short time by combining residual learning, skip architecture, and symmetric weight initialization.  ...  This paper shows that it is possible to train large and deep convolutional neural networks (CNN) for JPEG compression artifacts reduction, and that such networks can provide significantly better reconstruction  ...  TE01020415) and the Ministry of Education, Youth and Sports from the National Programme of Sustainability (NPU II) project IT4Innovations excellence in science (no. LQ1602).  ... 
arXiv:1605.00366v1 fatcat:fjsdzrl2rvafrhkzg7273b3lcu

A Survey on Deep Convolutional Neural Networks for Image Steganography and Steganalysis

2020 KSII Transactions on Internet and Information Systems  
In this paper, we analyzed current research states from the latest image steganography and steganalysis frameworks based on deep learning.  ...  Our objective is to provide for future researchers the work being done on deep learning-based image steganography & steganalysis and highlights the strengths and weakness of existing up-to-date techniques  ...  In the following section we will cover several deep learning algorithms that hide information in jpeg domain & will also highlight pros and cons. In 2017, Zeng et al.  ... 
doi:10.3837/tiis.2020.03.017 fatcat:7ci7bfbjsfd2nn5yagnv2h3ora

Fast object detection in compressed JPEG Images [article]

Benjamin Deguerre, Clément Chatelain, Gilles Gasso
2019 arXiv   pre-print
Most of these deep learning models rely on RGB images to localize and identify objects in the image.  ...  To alleviate this drawback, we propose a fast deep architecture for object detection in JPEG images, one of the most widespread compression format.  ...  We thank ACTEMIUM Paris Transport for the dataset and the funding. We thank CRIANN for the GPU computation facilities.  ... 
arXiv:1904.08408v1 fatcat:tons6t6aizhppj6h2yrenlyadu

Large-Scale JPEG Image Steganalysis Using Hybrid Deep-Learning Framework

Jishen Zeng, Shunquan Tan, Bin Li, Jiwu Huang
2018 IEEE Transactions on Information Forensics and Security  
In this paper we propose a generic hybrid deep-learning framework for JPEG steganalysis incorporating the domain knowledge behind rich steganalytic models.  ...  The second stage is a compound deep neural network containing multiple deep subnets in which the model parameters are learned in the training procedure.  ...  Acknowledgment The authors would like to thank DDE Laboratory in SUNY Binghamton for sharing the source code of their steganalysis models online (http://dde.binghamton.edu/download/).  ... 
doi:10.1109/tifs.2017.2779446 fatcat:wofzsvwzorhkliovxur3epth6q

Local-Source Enhanced Residual Network for Steganalysis of Digital Images

Wonhyuk Ahn, Haneol Jang, Seung-Hun Nam, In-Jae Yu, Heung-Kyu Lee
2020 IEEE Access  
Moreover, the LSER exhibits state-of-the-art results compared with the existing work in both spatial and JPEG domain steganalysis.  ...  In this paper, we propose a local-source enhanced residual network (LSER) with end-to-end learning. The LSER is simple in its architecture but has two distinct characteristics from previous methods.  ...  In addition, domain knowledge is often incorporated in CNNs, especially for JPEG steganalysis. Xu [12] used 4 × 4 DCT basis functions to transform images from the pixel domain into the DCT domain.  ... 
doi:10.1109/access.2020.3011752 fatcat:33s4qxz6b5go5ozjpswwkniry4

CALPA-NET: Channel-pruning-assisted Deep Residual Network for Steganalysis of Digital Images [article]

Shunquan Tan, Weilong Wu, Zilong Shao, Qiushi Li, Bin Li, Jiwu Huang
2020 arXiv   pre-print
In this paper we propose CALPA-NET, a ChAnneL-Pruning-Assisted deep residual network architecture search approach to shrink the network structure of existing vast, over-parameterized deep-learning based  ...  Over the past few years, detection performance improvements of deep-learning based steganalyzers have been usually achieved through structure expansion.  ...  In JPEG domain, Chen et al. proposed a specific deep-learning based steganalyzer aware of JPEG phase [23] . Xu proposed a novel 20-layer framework with residual connections (named XuNet2) [24] .  ... 
arXiv:1911.04657v2 fatcat:fvj2tygabrgxnoqljv7ti2ucbi

An extended hybrid image compression based on soft-to-hard quantification

Haisheng Fu, Feng Liang, Bo Lei
2020 IEEE Access  
Recently, the deep learning methods have been widely used in lossy compression schemes, greatly improving image compression performance.  ...  including BPG in SSIM metric across a wide range of bit rates, when the images are coded in the RGB444 domain.  ...  JPEG, JPEG2000 and WebP codecs in the RGB444 domain.  ... 
doi:10.1109/access.2020.2994393 fatcat:3njlexiuzfbjxgsfjlrgmj7upi

Towards super resolution in the compressed domain of learning-based image codecs

Evgeniy Upenik, Michela Testolina, Touradj Ebrahimi, Andrew G. Tescher, Touradj Ebrahimi
2021 Applications of Digital Image Processing XLIV  
Unlike the traditional approaches to image compression, learning-based codecs exploit deep neural networks for reducing dimensionality of the input at the stage where a linear transform would be typically  ...  The signal representation after this stage, called latent space, carries the information in such a way that it can be interpreted by other deep neural networks without the need of decoding it.  ...  Unlike the traditional approaches for image compression, learning-based codecs exploit deep neural networks for reducing dimensionality of the input at the stage where a linear transform would be typically  ... 
doi:10.1117/12.2597833 fatcat:xmp6ajpmdbb7to4pyn3bfd6eae

D^3: Deep Dual-Domain Based Fast Restoration of JPEG-Compressed Images [article]

Zhangyang Wang, Ding Liu, Shiyu Chang, Qing Ling, Yingzhen Yang, and Thomas S. Huang
2016 arXiv   pre-print
In this paper, we design a Deep Dual-Domain (D^3) based fast restoration model to remove artifacts of JPEG compressed images.  ...  It leverages the large learning capacity of deep networks, as well as the problem-specific expertise that was hardly incorporated in the past design of deep architectures.  ...  While the spatial redundancies in the pixel domain were exploited by a learned dictionary [2] , the residual redundancies in the DCT domain were also utilized to directly restore DCT coefficients.  ... 
arXiv:1601.04149v3 fatcat:56v7f6otanbevnc26rrflu3kk4
« Previous Showing results 1 — 15 out of 1,794 results