伍家松

发布者:伍家松发布时间:2019-01-05浏览次数:14022

基本信息
姓 名:伍家松性 别:出生年月:请输入
学 位:博士职 称:副教授资 格:博士生导师
行政机构:影像科学与技术系研究方向:深度学习,信号与图像处理
电 话:83794249传 真:
电子邮件:jswu@seu.edu.cn
地 址:四牌楼2号邮 编:210096


详细信息

一、学习及工作经历
2021/07-至今,东南大学,计算机科学与工程学院影像科学与技术系,博士生导师

2020/05-至今, 东南大学, 计算机科学与工程学院影像科学与技术系,副教授

2019/03-至今,东南大学,计算机科学与工程学院影像科学与技术系,硕士生导师

2012/03-2020/04,东南大学,计算机科学与工程学院影像科学与技术系,讲师

2009/07–2012/03, 雷恩第一大学, 信号处理与通信, 博士(中法联合培养),导师:Lotfi Senhadji教授
2007/03–2012/11, 东南大学, 生物医学工程, 博士, 导师:舒华忠教授
2005/09–2007/03, 东南大学, 生物医学工程, 硕士研究生(提前攻博),导师:於文雪副教授
2001/09–2005/06, 南华大学, 生物医学工程, 学士, 导师:赵修良教授


二、获奖

2021年东南大学教学成果奖研究生教育二等奖(排名第4)

2019年海洋工程科学技术奖二等奖1项(排名第4)

2018年江苏省教育科学研究成果奖三等奖1项(排名第2)

2017年国家科技进步二等奖1项(排名第13)
2012年获教育部自然科学二等奖1项(排名第3)
2010年中国国家留学基金委“国家优秀自费留学生奖学金”
2009年法国外交部“艾菲尔(Eiffel)博士奖学金”


三、承担项目

国家自然科学基金青年项目1项《基于压缩感知,矩阵填充和鲁棒的主成分分析的四元数信号处理方法研究》(2013-2015)

国家自然科学基金面上项目1项《复数及四元数域卷积神经网络的构造方法及其应用研究》(2019-2022)

装备预研共用技术项目课题1项《语音伪造和防伪、多模态音视频联合检测》(2022-2024)


四、论文发表和专利申请

研究方向:深度学习,卷积网络,图像和视频描述,自监督学习,离散正交变换等。自2008年以来,发表学术论文71篇,其中SCI论文39篇,包括8篇IEEE Transactions 系列论文。申请专利18项,目前已经授权13项。

Google Scholar主页:  https://scholar.google.com/citations?hl=zh-CN&user=mc7pLzwAAAAJ


4.1研究方向

   4.1.1 卷积网络的构造

        1. 分数阶小波散射网络构造了一种新的结构简单的分数阶小波散射网络,该网络是将复数小波(包括Morlet小波与双数复数小波)扩展到分数阶域并根据散射网络的结构构建而成,证明了网络的能量收敛性。该网络在MICCAI 2015 腺体分割任务中比传统的小波散射网络具有更好的性能,甚至与结构复杂的多层卷积神经网络具有相当地性能。成果发表于生物医学工程领域权威期刊《IEEE Transactions on Biomedical Engineering》【10】。

        2. 主成分分析网络的解释与推广K-L变换(Karhunen–Loève Transform)在模式识别领域又被称为主成分分析 (Principal Component Analysis: PCA),是一种经典的特征提取方法。可解释性人工智能(Artificial Intelligence: AI)是当前深度学习研究领域的热点之一。研究组从图像能量的角度对主成分分析网络(Principal Component Analysis Network: PCANet)进行了解释,论文发表在人工智能领域知名期刊《Neurocomputing》【12】。在此基础上,为了更有效的处理彩色图像,研究组将PCANet扩展到四元数域,提出了四元数主成分分析网络(Quaternion PCANet: QPCANet),成果发表于人工智能领域知名期刊《Neurocomputing》【21】。为了更有效的处理任意维度的张量类型数据,研究组提出了多线性主成分分析网络(Multilinear PCANet: MPCANet),成果发表于期刊《IEEE Access》【17】。     

       3. 四元数Softmax分类器。构造了一种新的四元数Softmax分类器,使之能够对四元数特征(比如: 彩色图像的四元数表达矩阵、四元数主成分分析特征等)进行正确的分类。成果发表于信号处理领域知名期刊《Electronic Letters》【24】。

       4. 八元数卷积神经网络的构造与应用。深度学习是机器学习方法和应用领域中的一个热门研究主题。实值神经网络(Real NNs),尤其是深度实数网络(DRNs)已广泛用于许多研究领域。近年来,深度复数网络(DCN)和深度四元数网络(DQN)引起了越来越多的关注。八元数代数是复数代数和四元数代数的扩展,可以提供更有效和紧凑的表达式。研究组构建了一个深度八元数网络(DON)的通用框架,并提供了诸如八元数卷积,八元数批量归一化和八元数权重初始化等DON的主要构建模块。然后,将DON用于CIFAR-10和CIFAR-100数据集的图像分类任务中。与DRN,DCN和DQN相比,提出的DON具有更好的收敛性和更高的分类精度。DON的成功还可以通过多任务学习来解释。论文发表在《Neurocomputing 》杂志【1】。


   4.1.2 图像描述和视频描述

        4.1.2.1 图像描述(Image Captioning)

        图像描述生成(Image Captioning)是一个融合计算机视觉、自然语言处理和机器学习的综合问题,它类似于翻译一幅图像为一段描述文字("看图说话")。该任务对于人类来说非常容易,但是对于机器却非常具有挑战性,它不仅需要利用模型去理解图片的内容并且还需要用自然语言去表达它们之间的关系。除此之外,模型还需要能够抓住图像的语义信息,并且生成人类可读的句子。

       直观的感知图像描述技术可以观看李飞飞教授的“如何教计算机理解图片”视频!


        4.1.2.2 视频描述(Video Captioning)

         视频描述是计算机对视频生成一段文字描述,这对在线的视频的检索等有很大帮助。不同于图像这种静态的空间信息,视频除了空间信息还包括时序信息,同时还有声音信息,这就表示一段视频比图像包含的信息更多,同时要求提取的特征也就更多,这对生成一段准确的文字描述是重大的挑战。

         具体可以参考:  

Xu J, Mei T, Yao T, et al. MSR-VTT: A large video description dataset for bridging video and language[C]//Proceedings of the IEEE conference on computer vision and pattern recognition. 2016: 5288-5296.

S. Venugopalan, M. Rohrbach, J. Donahue, R. J. Mooney, T. Darrell, and K. Saenko. Sequence to sequence - video to text. In ICCV, 2015.


   4.1.3 语音处理

        4.1.3.1 语音分离/鸡尾酒会问题(Speech separation/Cocktail Party Problem)

        多说话人语音分离技术,从广义上讲是从一段包含多人说话声及噪声的音频信号中把不同类别的声音信号区分开来,以在嘈杂的环境中关注指定语音的技术。语音分离技术对于推动未来人机交互有着不可或缺的奠基作用。

        直观的感知语音分离技术可以观看谷歌的“谷歌最新黑科技,能分离两个人声视频!


        4.1.3.2 语音合成(Speech Synthesis/ Text-To-Speech)

        语音合成即将文字用语音发出来,实现“让机器说话”。2016年谷歌发布语音合成网络WaveNet开始,深度学习在语音合成技术上的研究迅速发展,至今提出了许多新的模型和技术。

        比如:克隆你的声音项目,“由声音生成人脸”语音画像项目,“由声音生成动作”项目等。


4.2 SCI论文列表

Arxiv:

[1]Qingchun Li, Jiasong Wu*, Yilun Kong, Youyong Kong, Guanyu Yang, Lotfi Senhadji, Huazhong Shu. Speech denoising using only single noisy audio samples. 2021.

[2]Fuzhi Wu#, Jiasong Wu#, Youyong Kong, Chunfeng Yang, Guanyu Yang, Huazhong Shu, Lotfi Senhadji. Modulation theory: A new entry point for convolutional neural networks. 2021.

[3]Jiasong Wu*, Taotao Li, Youyong Kong, Guanyu Yang, Lotfi Senhadji, Huazhong Shu. SLNSpeech: solving extended speech separation problem by the help of sign language. 2020. arXiv: 2007.10629.


[4]Jiasong Wu*, Fuzhi Wu, Jieyuan Liu, Youyong Kong, Xu Han, Lotfi Senhadji, Huazhong Shu. Phase-only signal reconstruction by MagnitudeCut. arxiv: 1603.00210.

已发表:

[1]Jiasong Wu*, Xiang Qiu, Jing Zhang, Fuzhi Wu, Youyong Kong, Guanyu Yang, Lotfi Senhadji, Huazhong Shu. Fractional wavelet based generative scattering networks. Frontiers in Neurorobotics. 2021 

[2]Xilin Liu , Yongfei Wu , Hao Zhang , Jiasong Wu , Liming Zhang. Quaternion discrete fractional Krawtchouk transform and its application in color image encryption and watermarking. Signal Processing. 189: 108275 (2021)

[3]Yan Zhang, Yifei Li, Youyong Kong, Jiasong Wu, Jian Yang, Huazhong Shu, Gouenou Coatrieux. GSCFN: A graph self-construction and fusion network for  semisupervised brain tissue segmentation in MRI. Neurocomputing, vol. 455, pp. 23-37, 2021.

[4]Yuting He, Guanyu Yang, Jian Yang, Yang Chen, Youyong Kong, Jiasong Wu, Lijun Tang, Xiaomei Zhu, Jean-Louis Dillenseger, Pengfei Shao, Shaobo Zhang, Huazhong Shu, Jean-Louis Coatrieux, Shuo Li. Dense biased networks with deep priori anatomy and hard region adaptation: Semi-supervised learning for fine renal artery segmentation. Medical Image Anal. 63: 101722 (2020)

[5]J.S. Wu*, L. Xu, F.Z. Wu, Y.Y. Kong, L. Senhadji, H.Z. Shu. Deep octonion networks. Neurocomputing, vol. 397, pp. 179-191, 2020.

[6]J.S. Wu*, F.Z. Wu, Q.H. Yang, Y. Zhang, X.L. Liu, Y.Y. Kong, L. Senhadji, H.Z. Shu, Fractional spectral graph wavelets and their applications. Mathematical Problems in Engineering. Vol. 2020, Article ID 2568179, 18 Pages. 

[7]Li Liu, Da Chen, Laurent D. Cohen, Jiasong Wu, Michel Paques, Huazhong Shu*, Anisotropic tubular minimal path model with fast marching front freezing scheme, Pattern Recognition, vol. 104, 2020, doi: https://doi.org/10.1016/j.patcog.2020.107349

[8] Aodong Shen, Han Dong, Kun Wang, Youyong Kong, Jiasong Wu, Huazhong Shu. Automatic extraction of blur regions on a single image based on semantic segmentation. IEEE Access, vol. 8, pp. 44867- 44878, 2020.

[9]Peng Yang, Guowei Yang, Xun Gong, Pingping Wu, Xu Han, Jiasong Wu, Caisen Chen. Instance segmentation network with self-distillation for scene text detection. IEEE Access, vol. 8, pp. 45825 – 45836, 2020.

[10]Y. Zhang, J. S. Wu, Y.Y. Kong, G. Coatrieux, H.Z. Shu. Image denoising via a non-local patch graph total variation. PLOS ONE, vol. 14, no. 12, pp. 1-16, 2019.

[11]Y. Y. Kong, J.S. Wu, G.Y. Yang, Y.L. Zuo, Y. Chen, H.Z. Shu, J.L. Coatrieux. Iterative spatial fuzzy clustering for 3D brain magnetic resonance image supervoxel segmentation. Journal of Neuroscience Methods, vol. 311, pp. 17-29, 2019.

[12]R.J. Ge, G.Y. Yang, J.S. Wu, Y. Chen*, G. Coatrieux, L.M. Luo. A novel chaos-based symmetric image encryption using bit-pair level process. IEEE Access, vol. 7, no. 1, pp. 99470-99480, 2019.

[13]X.L. Liu*, Y.F. Wu, Z.H. Shao, J.S. Wu. The modified generic polar harmonic transforms for image representation. Pattern Analysis and Applications. https://doi.org/10.1007/s10044-019-00840-0

[14]L. Liu, J. S. Wu, D. W. Li, L. Senhadji, H. Z. Shu*. Fractional wavelet scattering network and applications. IEEE Transactions on Biomedical Engineering, vol. 66, no. 2, pp. 553-563, 2019.

[15]J.S. Wu*, F.Z. Wu, Z.F. Dong, K.W. Song, Y.Y. Kong, L. Senhadji, H.Z. Shu. Fast gray code kernel algorithm for the sliding conjugate symmetric sequency-ordered complex Hadamard transform. IEEE Access, vol. 6, no. 1, 2018: 56029-56045.

[16]J. S. Wu*, S. J. Qiu, Y. Y. Kong, L. Y. Jiang, Y. Chen, W. K. Yang, L. Senhadji, H. Z. Shu. PCANet: An energy perspective. Neurocomputing, vol. 313, pp. 271-287, 2018.

[17]Xilin Liu, Yongfei Wu, Zhuhong Shao, Jiasong Wu, Huazhong Shu. Color image watermarking using a discrete trinion Fourier transform. Journal of Electronic Imaging, 2018, 27 (4):043046-1-14.

[18]L. Y. Jiang, R.G. He, J. Liu, Y. Chen, J.S. Wu, H.Z. Shu. Phase-constrained parallel magnetic resonance imaging reconstruction based on low-rank matrix completion. IEEE Access, vol. 6, pp. 4941-4954, 2018.

[19]L.Y. Jiang, R.G. He, Y.P. Hong, J.S. Wu, H.Z. Shu. Two-dimensional active raypath separation using examination of the roots of the spectrum polynomial. The Journal of the Acoustical Society of America, vol. 142, no. 4, pp. EL408-EL414, 2017.

[20]W. Yang, H. Zhang, J. Yang, J. Wu, X. Yin, Y. Chen, H. Shu, L. Luo, G. Coatrieux, Z. Gui, Q. Feng. “Improving low-dose CT image using residual convolutional network,” IEEE Access, vol. 5, pp. 24698-24705, 2017.

[21]Jiasong Wu*, Shijie Qiu, Rui Zeng, Youyong Kong, Lotfi Senhadji, Huazhong Shu. Multilinear principal component analysis network for tensor object classification. IEEE Access, vol. 5, pp. 3322-3331, 2017.

[22]Xilin Liu, Guoniu Han, Jiasong Wu, Zhuhong Shao, Gouenou Coatrieux, and Huazhong Shu*. Fractional Krawtchouk transform with an application to image watermarking. IEEE Transactions on Signal Processing, vol. 65, no. 7, pp. 1894-1908, 2017.

[23]Zhang J, Wu J, Coatrieux J L, et al. A Correlation Based Strategy for the Acceleration of Nonlocal Means Filtering Algorithm. Mathematical Problems in Engineering, pp. 1-7, 2016.

[24]Zhuhong Shao, Yuanyuan Shang, Rui Zeng, Huazhong Shu, Gouenou Coatrieux, Jiasong Wu.  Robust watermarking scheme for color image based on quaternion-type moment invariants and visual cryptography. Signal Processing: Image Communication. Volume 48, October 2016, Pages 12–21.

[25]Zeng R, Wu J S*, Shao Z H, Chen Y, Senhadji L, Shu H Z. Color image classification via quaternion principal component analysis network.  Neurocomputing, vol. 216, pp. 416-428, 2016.

[26]  B.J. Chen, G. Coatrieux, J.S. Wu, Z.F. Dong, J.L. Coatrieux, H.Z. Shu. Fast computation of sliding discrete Tchebichef moments and its application in duplicated regions detection. IEEE Trans. Signal Processing, vol. 63, no. 20, pp. 5424-5436, 2015.

[27]  Z. H. Shao, Y. P. Duan, G. Coatrieux, J. S. Wu, J. Y. Meng, H. Z. Shu. Combining double random phase encoding for color image watermarking in quaternion gyrator domain. Optics Communications, vol. 343, pp. 56-65, 2015.
[28]  R. Zeng, J.S. Wu*, Z.H. Shao, L. Senhadji, H.Z. Shu. Quaternion softmax classifier. IET Electronics Letters, vol. 50, no. 25, pp. 1929-1930, 2014.
[29]  F. Liao, J. L. Coatrieux, J.S. Wu, Huazhong Shu. A new fast algorithm for constrained four-directioal total variation image denoising problem. Mathematical Problems in Engineering, 2014, Article ID 815132, pp. 1-11.
[30]  X. Han, J.S. Wu, L. Wang, Y. Chen, L. Senhadji, H.Z. Shu. Linear Total Variation Approximate Regularized Nuclear Norm Optimization for Matrix Completion, Abstract and Applied Analysis, 2014, Article ID 765782, pp. 1-8.
[31] Y. Chen, H.Z. Shu, J. Yang, J.S. Wu, L.M. Luo, J. L. Coatrieux. 2-D Impulse noise suppression by recursive Gaussian maximum likelihood estimation. PLoS ONE, e96386,  vol. 9, no. 5, pp. 1-14, 2014.
[32] Z.H. Shao, H.Z. Shu, J.S. Wu, Z.F. Dong, G. Coatrieux, J. L. Coatrieux. Double color image encryption using iterative phase retrieval algorithm in quaternion gyrator domain. Optics Expresses, vol. 22, no. 5, pp. 4932-4942, 2014.
[33] Z.H. Shao, H.Z. Shu, J.S. Wu, B.J. Chen, J. L. Coatrieux. Quaternion Bessel-Fourier moments and their invariant descriptors for object reconstruction and recognition. Pattern Recognition, vol. 47, no. 2, pp. 603-611, Feb. 2014.
[34] J. Wang, S. J. Wang, Y. Chen, J.S. Wu, J. L. Coatrieux, and L. M. Luo. Metal artifact reduction in CT using fusion based prior image. Med. Phys. 2013, 40 (8), 081903-1-8.
[35] J.S. Wu, L. Wang, G. Y. Yang, L. Senhadji, L. M. Luo, H. Z. Shu. Sliding conjugate symmetric sequency-ordered complex Hadamard transform: fast algorithm and applications. IEEE Trans Circuits Syst-I: Regular Papers, 2012, 59 (6): 1321-1334.
[36] H. Z. Shu, J.S. Wu, C. F. Yang, L. Senhadji. Fast radix-3 algorithm for the generalized discrete Hartley transform of type II. IEEE Signal Process Lett, 2012, 19(6): 348-351.
[37] B. Wang, J. S. Wu, H. Z. Shu, L. M. Luo,  Shape description using sequency-ordered complex Hadamard transform. Optics Commun., 2011, 284(12): 2726-2729.
[38] J.S. Wu, H.Z. Shu, L. Wang, L. Senhadji. Fast algorithms for the computation of sliding sequency-ordered complex Hadamard transform. IEEE Trans. Signal Process., 2010, 58(11): 5901-5909.
[39] L.Y. Jiang, H.Z. Shu, J.S. Wu, L. Wang, L. Senhadji. A novel split-radix fast algorithm for 2-D discrete Hartley transform. IEEE Trans. Circuits Syst.-I, 2010, 57(4): 911-924.
[40] J.S. Wu, H.Z. Shu, L. Senhadji, L.M. Luo. Mixed-radix algorithm for the computation of forward and inverse MDCTs. IEEE Trans. Circuits Syst.-I, 2009, 56(4): 784-794.
[41] H.Z. Shu, J.S. Wu, L. Senhadji, L.M. Luo. New fast algorithm for modulated complex lapped transform with sine windowing function. IEEE Signal Process. Lett., 2009, 16(2): 93-96.
[42] J.S Wu, H.Z. Shu, L. Senhadji, L.M. Luo. Radix-3×3 algorithm for the 2-D discrete Hartley transform. IEEE Trans. Circuits Syst. -II, 2008, 55(6): 566-570.
[43] J.S. Wu, H.Z. Shu, L. Senhadji, L.M. Luo. A fast algorithm for the computation of 2-D forward and inverse MDCT. Signal Process., 2008, 88(6): 1436-1446.
[44] H.Z. Shu, J.S. Wu, L. Senhadji, L.M. Luo. Radix-2 algorithm for the fast computation of type-III 3-D discrete W transform. Signal Process., 2008, 88(1): 210-215.


五、所教课程

5.1 正在承担

计软智专业大三本科生课程《深度学习与应用(研讨)》、《深度学习课程设计》(2021年-至今)

选修课《面向司法领域的人工智能》(2019-至今)

东南大学苏州软件学院研究生课程《机器学习》(2017-至今)

东南大学计算机学院研究生课程《信号与图像处理新技术进展》(2016-至今)


5.2 曾经承担

软件专业大三本科生课程《运筹学》(2016-2020)、《深度学习导论》(2016-2020)

人工智能专业大三本科生课程《计算机视觉》(2021)

计算机专业大二本科生课程《信号与系统》(2019-2020):https://www.bilibili.com/video/BV1Nq4y1j7nj


六、人才培养

本人可以指导博士后;今年拟招收博士研究生(南京)1名,硕士研究生(南京)3-4名,硕士研究生(苏州)2-3名!欢迎各位保研和考研的同学与我联系!

培养硕士研究生26,在读17名。

6.1 在读研究生

2021级:吴婷婷、李萱、靳鸿祥、李康康、马瑛瑶、薛一帆、姜大朗

2020级:李清淳、孙域、殷哲贤、郭宗辉、邱祥、宋佳朋、施龙斌、张家伟

2019级:陈曦、王晨琳、文智奕、孟凡满、姜海峰、孙威、李静波、王正青、王辉


6.2 毕业研究生

毕业时间姓名:就业单位(QQ联系方式)
2014韩旭:法国雷恩第一大学读博(229897099);严路:腾讯上海(296628981);葛迦:浙江大学附属第一医院(371262769
2015曾瑞:昆士兰科技大学读博(376338108);冯永:南方科技大学工作(168132430);刘洁媛:华为南京(1095486276
2016吴丹:欢聚时代广州(1053531933
2017达臻:美团上海(1527614075);郑爱宇:兵器集团第207所(550842909
2018邱诗洁:滴滴北京(1063392760);魏黎明:唯品会上海(949156461
2019应雨婷:华为南京(932239167);任虹珊:华为南京(527890478);杨启晗:华为南京(392011109);朱小贝:华为南京(773139709
2020徐玲:国网铜陵(1754719936);夏金鹏:阿里(467715384);祝木林:创新奇智(1142732109);曹国栋:电信云(411324590);张景:中兴(913959805);申泽宇:读博(526462505
2021李晓燕:华为南京(785756304);  闻婷:虾皮(1014642494);  李桃桃:美团上海(1913278504); 李蒙:快手(416573420); 何东升:新华三(804583359