• 2017

  • SEP-Nets: Simple and Effective Pattern Networks [PDF]

    Zhe Li, Xiaoyu Wang, Xutao Lv, Tianbao Yang

    arXiv:1706.03912,

    RED-Net: A Recurrent Encoder-Decoder Network for Video-based Face Alignment [PDF]

    Xi Peng, Rogerio S. Feris, Xiaoyu Wang, Dimitris N. Metaxas

    IJCV, (Submitted to IJCV 2017)

    IJCV Special Issue on "Best Papers in ECCV"

    Exploring Personalized Neural Conversational Models [PDF]

    Satwik Kottur, Vitor Carvalho, Xiaoyu Wang

    Internation Conference on Artificial Intelligence, (IJCAI 2017)

    Deep Reinforcement Learning-based Image Captioning with Embedding Reward [PDF]

    Zhou Ren, Xiaoyu Wang, Ning Zhang, Xutao Lv, Jia Li

    Computer Vision and Pattern Recognition, (CVPR 2017)

    Oral Presentation

  • 2016

  • A Recurrent Encoder-Decoder Network for Sequential Face Alignment [PDF]

    Xi Peng, Rogerio S. Feris, Xiaoyu Wang, Dimitris N. Metaxas

    Computer Vision and Pattern Recognition, (ECCV 2016)

    Oral Presentation, invited to IJCV Special Issue on "Best Papers in ECCV"

  • 2015

  • Semantic-aware co-indexing for image retrieval [PDF]

    Shiliang Zhang, Ming Yang, Xiaoyu Wang, Yuanqing Lin

    IEEE Transactions on Pattern Analysis and Machine Intelligence, (T-PAMI 2015)

    Regionlets for Generic Object Detection [PDF] [Video] [Oral presentation] [ImageNet presentation]

    Xiaoyu Wang, Ming Yang, Shenghuo Zhu, Yuanqing Lin

    IEEE Transactions on Pattern Analysis and Machine Intelligence, (T-PAMI 2015)

    Check our longer version for more details about the Regionlets technology

    Hyper-class Augmented and Regularized Deep Learning for Fine-grained Image Classification [PDF]

    Saining Xie, Tianbao Yang, Xiaoyu Wang, Yuanqing Lin

    CVPR 2015, (CVPR 2015)

    Scalable feature matching by dual cascaded scalar quantization for image retrieval

    Wengang Zhou, Ming Yang, Xiaoyu Wang, Hongqiang Li, Yuanqing Lin, Qi Tian

    IEEE Transactions on Pattern Analysis and Machine Intelligence, (T-PAMI 2015)

    Cross Indexing with Grouplets

    Shiliang Zhang, Xiaoyu Wang, Yuanqing Lin, Qi Tian

    IEEE Transactions on Multimedia, (T-MM 2015)

  • 2014

  • Accurate Object Detection with Location Relaxation and Regionlets Relocalization [PDF]

    Chengjiang Long, Xiaoyu Wang, Gang Hua, Ming Yang, Yuanqing Lin

    in Proc. 12th Asian Conf. on Computer Vision(ACCV'2014), (ACCV 2014)

    Intern project under my supervision, it achieves the best performnce on PASCAL VOC 2007 as well as KITTI object deteciton dataset. Also check the tracking submissions on KITTI which utilize Regionlets to get state-of-art performace.

    Towards Codebook-free: Scalable Cascaded Hashing for Mobile Image Search [PDF]

    Wengang Zhou, Ming Yang, Houqiang Li, Xiaoyu Wang, Yuanqing Lin, Qi Tian

    in IEEE Transaction on Multimedia(T-MM' 2013), (T-MM 2014)

    Generic Object Detection with Dense Neural Patterns and Regionlets [PDF]

    Will Zou, Xiaoyu Wang, Miao Sun, Yuanqing Lin

    Technic Report, (TR 2013, BMVC 2014)

    Intern projects under my supervision, Regionlets with CNN, tractable computation.

  • 2013

  • Regionlets for Generic Object Detection [PDF] [Video] [Oral presentation] [ImageNet presentation]

    Xiaoyu Wang, Ming Yang, Shenghuo Zhu, Yuanqing Lin

    in Proc. IEEE 14th International Conf. on Computer Vision(ICCV'2013), (ICCV 2013)

    Oral presentation (acceptance rate < 2.5%), Runner up Winner in the ImageNet large scale object detection challenge (ILSVRC 2013)

    Semantic-aware Co-indexing for Near-duplicate Image Retrieval [PDF]

    Shiliang Zhang, Ming Yang, Xiaoyu Wang, Yuanqing Lin, Qi Tian

    in Proc. IEEE 14th International Conf. on Computer Vision(ICCV'2013), (ICCV 2013)

    Hierarchical Feature Pooling with Structure Learning: A new method for Pedestrain Detection [PDF]

    Xiaoyu Wang, Liangliang Cao, Rogerio Feris, Ankur Data, Tony X. Han

    in CVPR 2013 workshop, (CVPRW 2013)

    map the original feature to high dimensional space after coding and pooling. (We use the original feature instead of codes), can capture coase to fine spatial layout.

  • 2012

  • Histogram of Oriented Normal Vectors for Object Recognition with a Depth Sensor [PDF]

    Tang, X Wang, X Lv, TX Han, J Keller, Z He, M Skubic, S Lao

    The 11th Asian Conference on Computer Vision 2012(ACCV'2012), (ACCV 2012)

    Oral presentation (acceptance rate < 3.6%)

    Detection by Detections: Non-parametric Detector Adaptation for a Video [PDF]

    Xiaoyu Wang, Gang Hua, Tony X. Han

    in Proc. IEEE 14th International Conf. on Computer Vision(ICCV'2009), (CVPR 2012)

    A first try to use detections to help detections in a video.

  • 2011

  • Contextual Weighting for Vocabulary Tree Based Image Retrieval [PDF] [dataset]

    Xiaoyu Wang, Ming Yang, Timothee Cour, Shenghuo Zhu, Kai Yu, Tony X. Han

    in Proc. IEEE 13th International Conf. on Computer Vision(ICCV'2011), (ICCV 2011)

    Excellent performance with fast speed for large scale image retrieval.

    Efficient Re-Ranking in Vocabulary Tree Based Image Retrieval [PDF]

    Xiaoyu Wang, Ming Yang, Kai Yu

    in the 45th Asilomar Conference on Signals, Systems and Computers, (Asilomar 2011)

    Invited book chapter

  • 2010

  • Discriminative Tracking by Metric Learning [PDF]

    Xiaoyu Wang, Gang Hua, Tony X. Han

    in Proc. the 11th European Conf. on Computer Vision (ECCV'2010), (ECCV 2010)

    Discriminative Multiple Target Tracking [PDF]

    Xiaoyu Wang, Ming Yang, Kai Yu

    in Machine Learning for Vision based Motion Analysis, (MLVMA 2010)

    Invited book chapter

  • 2009

  • An HOG-LBP human detector with partial occlusion handling [PDF] [oral presentation] [code] [Video] [bibtex]

    Xiaoyu Wang, Tony X. Han, Shuicheng Yan

    in Proc. IEEE 14th International Conf. on Computer Vision(ICCV'2009), (ICCV 2009)

    Oral presentation (acceptance rate < 3.6%), ICCV 2009 most cited papers rank #1)

  • 2008

  • Build A Robot from Scratch with Steering Engines and Stepper Motors

    Xiaoyu Wang

    in Robotics Innovation, Science Press, 2008,

    Continue reading