|
|
|
题名
|
作者
|
年代
|
出处
|
被引量
|
| 1 | Limits of learning-based super-resolution algorithms显示文摘 | LIN Zhouchen HE Junfeng TANG Xiaoou | 2008 | International Journal of Computer Vision(S0920-5691)2008，80，1： | 1 |
| 2 | Ro- bust recovery of subspace structures by low-rank represen- tation显示文摘 | Liu Guangcan Lin Zhouchen Yan Shuicheng | 2013 | IEEE Transactions on Pattern Analysis and Ma- chine Intelligence2013，35，1： | 1 |
| 3 | Robust recovery ofsubspace structures by low-rank representation 显示文摘 | Liu Guangcan Lin Zhouchen Yan Shuicheng’ | 2013 | IEEE Trans onPattern Analysis and Machine Intelligence2013，35，1： | 1 |
| 4 | Fast, automatic and fine-grained tampered JPEG image detection via DCT co- efficient analysis显示文摘 | LIN Zhouchen HE Junfeng TANG Xiaoou | 2009 | Pattern Recognition2009，42，11： | 1 |
| 5 | Robust recovery of subspace structures by low-rank representation显示文摘 | LIU Guangcan LIN Zhouchen YAN Shuicheng | | 0，，01： | 1 |
| 6 | Fast,automatic and fine-grained tampered JPEG image detection via DCT coefficient anal-ysis显示文摘 | Lin Zhouchen He Junfeng | 2009 | Pattern Recognition2009，42，11： | 1 |
| 7 | A geometric method for optimal design of color filter arrays显示文摘 | HAO Pengwei LI Yan LIN Zhouchen | 2011 | IEEE Trans Image Process2011，20，3： | 1 |
| 8 | Robust recovery of subspace structures by low-rank representation显示文摘 | Liu Guangcan Lin Zhouchen Yan Shuicheng | 2013 | IEEE Transactions on Pattern Analysis and Machine Intelligence2013，35，1： | 1 |
| 9 | Ll-norm kernel discriminant analysis via Bayes error bound optimizatin for robust feature extraction 显示文摘 | HENG Wenming LIN Zhouchen WANG Haixian | 2014 | IEEE Transactions on Neural Networks and Learning Systems(S 1045-9227)2014，25，4： | 1 |
| 10 | Fundamental Limits of Reconstruction-Based Super-Resolution Alorithms Under Local Translation显示文摘 | Lin Zhouchen Shum Heung-Yeung | 2004 | IEEE Trans on Pattern Analysis and Machine Intelligence2004，26，1： | 1 |
| 11 | Robust recovery of subspace structures by low-rank representation显示文摘 | Liu Guangcan Lin Zhouchen Yan Shuicheng | 2013 | IEEE Trans on PAMI2013，35，1： | 1 |
| 12 | The aug- mented lagrange multiplier method for exact recovery of cor-rupted low-rank matrices 显示文摘 | Lin Zhouchen Chen Minming Wu Leqiu | 2010 | Mathematics : Optimization and Control2010，26，1： | 1 |
| 13 | Robust reco- very of subspace structures by low-rank representation 显示文摘 | Liu Guangcan Lin Zhouchen ~an Shuicheng | 2013 | IEEE Trans on Pattern Analysis and Machine Intelligence2013，，： | 1 |
| 14 | Multi-output Regression on the Output Manifold显示文摘 | Liu Guangcan Lin Zhouchen Yu Yong | 2009 | Pattern Recog- nition2009，42，： | 1 |
| 15 | Fast, automatic and fine-grained tampered JPEG image detection via DCT coefficient analysis显示文摘 | Zhouchen Lin Junfeng He Xiaoou Tang Chi-Keung Tang | 2009 | Pattern Recognition2009，，11： | 1 |
| 16 | Fast convex optimization algorithms for exact recovery of a corrupted low-rank matrix显示文摘 | LIN Zhouchen GANESH A WRIGHT J | 2009 | International Workshop on Computational Advances in Multi-Sensor Adaptive Processing(CAMSAP)2009，61，： | 1 |
| 17 | Robust recovery of subspace structures by low-rank representation显示文摘 | Liu Guangcan Lin Zhouchen Yan Shuicheng | 2012 | IEEE Transactions on Software Engineering2012，35，1： | 1 |
| 18 | 基于规则的连机英语字符笔迹标识的修正显示文摘近来,许多基于书写的方式使人们自然书写的数字笔迹输入成为可能。通常,书写时的污点或修改不仅使文本受污,而且看上去也不舒服,还影响到手写体的识别。本文首先论述笔迹修正的问题。我们提出了去除污点和修改笔迹的修正系统,使文本变得清晰、可识别,以改善手写体的识别率。基于规则的算法可处理大部分如单一笔划的同一笔划重复涂写、笔划间的中间部分重复涂写、改错、修改、插入以及书写顺序错误等情况。实验结果显示,该系统对笔迹标识的修正是有效的,并有希望改善其识别率。 | Zhouchen Lin Rongrong Wang Heung - Yeung Shum 李宏(译) | 2006 | 图象识别与自动化2006，，2： | 0 |
| 19 | Optimization-inspired manual architecture design and neural architecture search显示文摘Neural architecture has been a research focus in recent years due to its importance in deciding the performance of deep networks.Representative ones include a residual network(Res Net)with skip connections and a dense network(Dense Net)with dense connections.However,a theoretical guidance for manual architecture design and neural architecture search(NAS)is still lacking.In this paper,we propose a manual architecture design framework,which is inspired by optimization algorithms.It is based on the conjecture that an optimization algorithm with a good convergence rate may imply a neural architecture with good performance.Concretely,we prove under certain conditions that forward propagation in a deep neural network is equivalent to the iterative optimization procedure of the gradient descent algorithm minimizing a cost function.Inspired by this correspondence,we derive neural architectures from fast optimization algorithms,including the heavy ball algorithm and Nesterov’s accelerated gradient descent algorithm.Surprisingly,we find that we can deem the Res Net and Dense Net as special cases of the optimization-inspired architectures.These architectures offer not only theoretical guidance,but also good performances in image recognition on multiple datasets,including CIFAR-10,CIFAR-100,and Image Net.Moreover,we show that our method is also useful for NAS by offering a good initial search point or guiding the search space. | Yibo YANG Zhengyang SHEN Huan LI Zhouchen LIN | 2023 | Science China(Information Sciences)2023，66，11： | 0 |
| 20 | Newton design:designing CNNs with the family of Newton's methods显示文摘Nowadays,convolutional neural networks(CNNs)have led the developments of machine learning.However,most CNN architectures are obtained by manual design,which is empirical,time-consuming,and non-transparent.In this paper,we aim at offering better insight into CNN models from the perspective of optimization theory.We propose a unified framework for understanding and designing CNN architectures with the family of Newton’s methods,which is referred to as Newton design.Specifically,we observe that the standard feedforward CNN model(PlainNet)solves an optimization problem via a kind of quasi-Newton method.Interestingly,residual network(ResNet)can also be derived if we use a more general quasi-Newton method to solve this problem.Based on the above observations,we solve this problem via a better method,the Newton-conjugate-gradient(Newton-CG)method,which inspires Newton-CGNet.In the network design,we translate binary-value terms in the optimization schemes to dropout layers,so dropout modules naturally appear in the derived CNN structures with specific locations,rather than being an empirical training strategy.Extensive experiments on image classification and text categorization tasks verify that Newton-CGNets perform very competitively.Particularly,Newton-CGNets surpass their counterparts ResNets by over 4%on CIFAR-10 and over 10%on CIFAR-100,respectively. | Zhengyang SHEN Yibo YANG Qi SHE Changhu WANG Jinwen MA Zhouchen LIN | 2023 | Science China(Information Sciences)2023，66，6： | 0 |
您所在的机构暂未开通此部分的功能，请联系管理员