Alert button
Picture for Xiong Wang

Xiong Wang

Alert button

Fusing Monocular Images and Sparse IMU Signals for Real-time Human Motion Capture

Sep 01, 2023
Shaohua Pan, Qi Ma, Xinyu Yi, Weifeng Hu, Xiong Wang, Xingkang Zhou, Jijunnan Li, Feng Xu

Either RGB images or inertial signals have been used for the task of motion capture (mocap), but combining them together is a new and interesting topic. We believe that the combination is complementary and able to solve the inherent difficulties of using one modality input, including occlusions, extreme lighting/texture, and out-of-view for visual mocap and global drifts for inertial mocap. To this end, we propose a method that fuses monocular images and sparse IMUs for real-time human motion capture. Our method contains a dual coordinate strategy to fully explore the IMU signals with different goals in motion capture. To be specific, besides one branch transforming the IMU signals to the camera coordinate system to combine with the image information, there is another branch to learn from the IMU signals in the body root coordinate system to better estimate body poses. Furthermore, a hidden state feedback mechanism is proposed for both two branches to compensate for their own drawbacks in extreme input cases. Thus our method can easily switch between the two kinds of signals or combine them in different cases to achieve a robust mocap. %The two divided parts can help each other for better mocap results under different conditions. Quantitative and qualitative results demonstrate that by delicately designing the fusion method, our technique significantly outperforms the state-of-the-art vision, IMU, and combined methods on both global orientation and local pose estimation. Our codes are available for research at https://shaohua-pan.github.io/robustcap-page/.

* Accepted by SIGGRAPH ASIA 2023. Project page: https://shaohua-pan.github.io/robustcap-page/ 
Viaarxiv icon

DCCRN-KWS: an audio bias based model for noise robust small-footprint keyword spotting

May 23, 2023
Shubo Lv, Xiong Wang, Sining Sun, Long Ma, Lei Xie

Figure 1 for DCCRN-KWS: an audio bias based model for noise robust small-footprint keyword spotting
Figure 2 for DCCRN-KWS: an audio bias based model for noise robust small-footprint keyword spotting
Figure 3 for DCCRN-KWS: an audio bias based model for noise robust small-footprint keyword spotting
Figure 4 for DCCRN-KWS: an audio bias based model for noise robust small-footprint keyword spotting

Real-world complex acoustic environments especially the ones with a low signal-to-noise ratio (SNR) will bring tremendous challenges to a keyword spotting (KWS) system. Inspired by the recent advances of neural speech enhancement and context bias in speech recognition, we propose a robust audio context bias based DCCRN-KWS model to address this challenge. We form the whole architecture as a multi-task learning framework for both denosing and keyword spotting, where the DCCRN encoder is connected with the KWS model. Helped with the denoising task, we further introduce an audio context bias module to leverage the real keyword samples and bias the network to better iscriminate keywords in noisy conditions. Feature merge and complex context linear modules are also introduced to strength such discrimination and to effectively leverage contextual information respectively. Experiments on the internal challenging dataset and the HIMIYA public dataset show that our DCCRN-KWS system is superior in performance, while ablation study demonstrates the good design of the whole model.

* Accepted by INTERSPEECH2023 
Viaarxiv icon

Two Stage Contextual Word Filtering for Context bias in Unified Streaming and Non-streaming Transducer

Jan 17, 2023
Zhanheng Yang, Sining Sun, Xiong Wang, Yike Zhang, Long Ma, Lei Xie

Figure 1 for Two Stage Contextual Word Filtering for Context bias in Unified Streaming and Non-streaming Transducer
Figure 2 for Two Stage Contextual Word Filtering for Context bias in Unified Streaming and Non-streaming Transducer
Figure 3 for Two Stage Contextual Word Filtering for Context bias in Unified Streaming and Non-streaming Transducer
Figure 4 for Two Stage Contextual Word Filtering for Context bias in Unified Streaming and Non-streaming Transducer

It is difficult for an end-to-end (E2E) ASR system to recognize words such as named entities appearing infrequently in the training data. A widely used method to mitigate this issue is feeding contextual information into the acoustic model. A contextual word list is necessary, which lists all possible contextual word candidates. Previous works have proven that the size and quality of the list are crucial. A compact and accurate list can boost the performance significantly. In this paper, we propose an efficient approach to obtain a high quality contextual word list for a unified streaming and non-streaming based Conformer-Transducer (C-T) model. Specifically, we make use of the phone-level streaming output to first filter the predefined contextual word list. During the subsequent non-streaming inference, the words in the filtered list are regarded as contextual information fused into non-casual encoder and decoder to generate the final recognition results. Our approach can take advantage of streaming recognition hypothesis, improve the accuracy of the contextual ASR system and speed up the inference process as well. Experiments on two datasets demonstrates over 20% relative character error rate reduction (CERR) comparing to the baseline system. Meanwile, the RTF of our system can be stabilized within 0.15 when the size of the contextual word list grows over 6,000.

Viaarxiv icon

A Data-Adaptive Prior for Bayesian Learning of Kernels in Operators

Dec 29, 2022
Neil K. Chada, Quanjun Lang, Fei Lu, Xiong Wang

Figure 1 for A Data-Adaptive Prior for Bayesian Learning of Kernels in Operators
Figure 2 for A Data-Adaptive Prior for Bayesian Learning of Kernels in Operators
Figure 3 for A Data-Adaptive Prior for Bayesian Learning of Kernels in Operators
Figure 4 for A Data-Adaptive Prior for Bayesian Learning of Kernels in Operators

Kernels are efficient in representing nonlocal dependence and they are widely used to design operators between function spaces. Thus, learning kernels in operators from data is an inverse problem of general interest. Due to the nonlocal dependence, the inverse problem can be severely ill-posed with a data-dependent singular inversion operator. The Bayesian approach overcomes the ill-posedness through a non-degenerate prior. However, a fixed non-degenerate prior leads to a divergent posterior mean when the observation noise becomes small, if the data induces a perturbation in the eigenspace of zero eigenvalues of the inversion operator. We introduce a data-adaptive prior to achieve a stable posterior whose mean always has a small noise limit. The data-adaptive prior's covariance is the inversion operator with a hyper-parameter selected adaptive to data by the L-curve method. Furthermore, we provide a detailed analysis on the computational practice of the data-adaptive prior, and demonstrate it on Toeplitz matrices and integral operators. Numerical tests show that a fixed prior can lead to a divergent posterior mean in the presence of any of the four types of errors: discretization error, model error, partial observation and wrong noise assumption. In contrast, the data-adaptive prior always attains posterior means with small noise limits.

* 30 pages, 8 figures 
Viaarxiv icon

Biomedical image analysis competitions: The state of current participation practice

Dec 16, 2022
Matthias Eisenmann, Annika Reinke, Vivienn Weru, Minu Dietlinde Tizabi, Fabian Isensee, Tim J. Adler, Patrick Godau, Veronika Cheplygina, Michal Kozubek, Sharib Ali, Anubha Gupta, Jan Kybic, Alison Noble, Carlos Ortiz de Solórzano, Samiksha Pachade, Caroline Petitjean, Daniel Sage, Donglai Wei, Elizabeth Wilden, Deepak Alapatt, Vincent Andrearczyk, Ujjwal Baid, Spyridon Bakas, Niranjan Balu, Sophia Bano, Vivek Singh Bawa, Jorge Bernal, Sebastian Bodenstedt, Alessandro Casella, Jinwook Choi, Olivier Commowick, Marie Daum, Adrien Depeursinge, Reuben Dorent, Jan Egger, Hannah Eichhorn, Sandy Engelhardt, Melanie Ganz, Gabriel Girard, Lasse Hansen, Mattias Heinrich, Nicholas Heller, Alessa Hering, Arnaud Huaulmé, Hyunjeong Kim, Bennett Landman, Hongwei Bran Li, Jianning Li, Jun Ma, Anne Martel, Carlos Martín-Isla, Bjoern Menze, Chinedu Innocent Nwoye, Valentin Oreiller, Nicolas Padoy, Sarthak Pati, Kelly Payette, Carole Sudre, Kimberlin van Wijnen, Armine Vardazaryan, Tom Vercauteren, Martin Wagner, Chuanbo Wang, Moi Hoon Yap, Zeyun Yu, Chun Yuan, Maximilian Zenk, Aneeq Zia, David Zimmerer, Rina Bao, Chanyeol Choi, Andrew Cohen, Oleh Dzyubachyk, Adrian Galdran, Tianyuan Gan, Tianqi Guo, Pradyumna Gupta, Mahmood Haithami, Edward Ho, Ikbeom Jang, Zhili Li, Zhengbo Luo, Filip Lux, Sokratis Makrogiannis, Dominik Müller, Young-tack Oh, Subeen Pang, Constantin Pape, Gorkem Polat, Charlotte Rosalie Reed, Kanghyun Ryu, Tim Scherr, Vajira Thambawita, Haoyu Wang, Xinliang Wang, Kele Xu, Hung Yeh, Doyeob Yeo, Yixuan Yuan, Yan Zeng, Xin Zhao, Julian Abbing, Jannes Adam, Nagesh Adluru, Niklas Agethen, Salman Ahmed, Yasmina Al Khalil, Mireia Alenyà, Esa Alhoniemi, Chengyang An, Talha Anwar, Tewodros Weldebirhan Arega, Netanell Avisdris, Dogu Baran Aydogan, Yingbin Bai, Maria Baldeon Calisto, Berke Doga Basaran, Marcel Beetz, Cheng Bian, Hao Bian, Kevin Blansit, Louise Bloch, Robert Bohnsack, Sara Bosticardo, Jack Breen, Mikael Brudfors, Raphael Brüngel, Mariano Cabezas, Alberto Cacciola, Zhiwei Chen, Yucong Chen, Daniel Tianming Chen, Minjeong Cho, Min-Kook Choi, Chuantao Xie Chuantao Xie, Dana Cobzas, Julien Cohen-Adad, Jorge Corral Acero, Sujit Kumar Das, Marcela de Oliveira, Hanqiu Deng, Guiming Dong, Lars Doorenbos, Cory Efird, Di Fan, Mehdi Fatan Serj, Alexandre Fenneteau, Lucas Fidon, Patryk Filipiak, René Finzel, Nuno R. Freitas, Christoph M. Friedrich, Mitchell Fulton, Finn Gaida, Francesco Galati, Christoforos Galazis, Chang Hee Gan, Zheyao Gao, Shengbo Gao, Matej Gazda, Beerend Gerats, Neil Getty, Adam Gibicar, Ryan Gifford, Sajan Gohil, Maria Grammatikopoulou, Daniel Grzech, Orhun Güley, Timo Günnemann, Chunxu Guo, Sylvain Guy, Heonjin Ha, Luyi Han, Il Song Han, Ali Hatamizadeh, Tian He, Jimin Heo, Sebastian Hitziger, SeulGi Hong, SeungBum Hong, Rian Huang, Ziyan Huang, Markus Huellebrand, Stephan Huschauer, Mustaffa Hussain, Tomoo Inubushi, Ece Isik Polat, Mojtaba Jafaritadi, SeongHun Jeong, Bailiang Jian, Yuanhong Jiang, Zhifan Jiang, Yueming Jin, Smriti Joshi, Abdolrahim Kadkhodamohammadi, Reda Abdellah Kamraoui, Inha Kang, Junghwa Kang, Davood Karimi, April Khademi, Muhammad Irfan Khan, Suleiman A. Khan, Rishab Khantwal, Kwang-Ju Kim, Timothy Kline, Satoshi Kondo, Elina Kontio, Adrian Krenzer, Artem Kroviakov, Hugo Kuijf, Satyadwyoom Kumar, Francesco La Rosa, Abhi Lad, Doohee Lee, Minho Lee, Chiara Lena, Hao Li, Ling Li, Xingyu Li, Fuyuan Liao, KuanLun Liao, Arlindo Limede Oliveira, Chaonan Lin, Shan Lin, Akis Linardos, Marius George Linguraru, Han Liu, Tao Liu, Di Liu, Yanling Liu, João Lourenço-Silva, Jingpei Lu, Jiangshan Lu, Imanol Luengo, Christina B. Lund, Huan Minh Luu, Yi Lv, Yi Lv, Uzay Macar, Leon Maechler, Sina Mansour L., Kenji Marshall, Moona Mazher, Richard McKinley, Alfonso Medela, Felix Meissen, Mingyuan Meng, Dylan Miller, Seyed Hossein Mirjahanmardi, Arnab Mishra, Samir Mitha, Hassan Mohy-ud-Din, Tony Chi Wing Mok, Gowtham Krishnan Murugesan, Enamundram Naga Karthik, Sahil Nalawade, Jakub Nalepa, Mohamed Naser, Ramin Nateghi, Hammad Naveed, Quang-Minh Nguyen, Cuong Nguyen Quoc, Brennan Nichyporuk, Bruno Oliveira, David Owen, Jimut Bahan Pal, Junwen Pan, Wentao Pan, Winnie Pang, Bogyu Park, Vivek Pawar, Kamlesh Pawar, Michael Peven, Lena Philipp, Tomasz Pieciak, Szymon Plotka, Marcel Plutat, Fattaneh Pourakpour, Domen Preložnik, Kumaradevan Punithakumar, Abdul Qayyum, Sandro Queirós, Arman Rahmim, Salar Razavi, Jintao Ren, Mina Rezaei, Jonathan Adam Rico, ZunHyan Rieu, Markus Rink, Johannes Roth, Yusely Ruiz-Gonzalez, Numan Saeed, Anindo Saha, Mostafa Salem, Ricardo Sanchez-Matilla, Kurt Schilling, Wei Shao, Zhiqiang Shen, Ruize Shi, Pengcheng Shi, Daniel Sobotka, Théodore Soulier, Bella Specktor Fadida, Danail Stoyanov, Timothy Sum Hon Mun, Xiaowu Sun, Rong Tao, Franz Thaler, Antoine Théberge, Felix Thielke, Helena Torres, Kareem A. Wahid, Jiacheng Wang, YiFei Wang, Wei Wang, Xiong Wang, Jianhui Wen, Ning Wen, Marek Wodzinski, Ye Wu, Fangfang Xia, Tianqi Xiang, Chen Xiaofei, Lizhan Xu, Tingting Xue, Yuxuan Yang, Lin Yang, Kai Yao, Huifeng Yao, Amirsaeed Yazdani, Michael Yip, Hwanseung Yoo, Fereshteh Yousefirizi, Shunkai Yu, Lei Yu, Jonathan Zamora, Ramy Ashraf Zeineldin, Dewen Zeng, Jianpeng Zhang, Bokai Zhang, Jiapeng Zhang, Fan Zhang, Huahong Zhang, Zhongchen Zhao, Zixuan Zhao, Jiachen Zhao, Can Zhao, Qingshuo Zheng, Yuheng Zhi, Ziqi Zhou, Baosheng Zou, Klaus Maier-Hein, Paul F. Jäger, Annette Kopp-Schneider, Lena Maier-Hein

The number of international benchmarking competitions is steadily increasing in various fields of machine learning (ML) research and practice. So far, however, little is known about the common practice as well as bottlenecks faced by the community in tackling the research questions posed. To shed light on the status quo of algorithm development in the specific field of biomedical imaging analysis, we designed an international survey that was issued to all participants of challenges conducted in conjunction with the IEEE ISBI 2021 and MICCAI 2021 conferences (80 competitions in total). The survey covered participants' expertise and working environments, their chosen strategies, as well as algorithm characteristics. A median of 72% challenge participants took part in the survey. According to our results, knowledge exchange was the primary incentive (70%) for participation, while the reception of prize money played only a minor role (16%). While a median of 80 working hours was spent on method development, a large portion of participants stated that they did not have enough time for method development (32%). 25% perceived the infrastructure to be a bottleneck. Overall, 94% of all solutions were deep learning-based. Of these, 84% were based on standard architectures. 43% of the respondents reported that the data samples (e.g., images) were too large to be processed at once. This was most commonly addressed by patch-based training (69%), downsampling (37%), and solving 3D analysis tasks as a series of 2D tasks. K-fold cross-validation on the training set was performed by only 37% of the participants and only 50% of the participants performed ensembling based on multiple identical models (61%) or heterogeneous models (39%). 48% of the respondents applied postprocessing steps.

Viaarxiv icon

CaTT-KWS: A Multi-stage Customized Keyword Spotting Framework based on Cascaded Transducer-Transformer

Jul 04, 2022
Zhanheng Yang, Sining Sun, Jin Li, Xiaoming Zhang, Xiong Wang, Long Ma, Lei Xie

Figure 1 for CaTT-KWS: A Multi-stage Customized Keyword Spotting Framework based on Cascaded Transducer-Transformer
Figure 2 for CaTT-KWS: A Multi-stage Customized Keyword Spotting Framework based on Cascaded Transducer-Transformer
Figure 3 for CaTT-KWS: A Multi-stage Customized Keyword Spotting Framework based on Cascaded Transducer-Transformer
Figure 4 for CaTT-KWS: A Multi-stage Customized Keyword Spotting Framework based on Cascaded Transducer-Transformer

Customized keyword spotting (KWS) has great potential to be deployed on edge devices to achieve hands-free user experience. However, in real applications, false alarm (FA) would be a serious problem for spotting dozens or even hundreds of keywords, which drastically affects user experience. To solve this problem, in this paper, we leverage the recent advances in transducer and transformer based acoustic models and propose a new multi-stage customized KWS framework named Cascaded Transducer-Transformer KWS (CaTT-KWS), which includes a transducer based keyword detector, a frame-level phone predictor based force alignment module and a transformer based decoder. Specifically, the streaming transducer module is used to spot keyword candidates in audio stream. Then force alignment is implemented using the phone posteriors predicted by the phone predictor to finish the first stage keyword verification and refine the time boundaries of keyword. Finally, the transformer decoder further verifies the triggered keyword. Our proposed CaTT-KWS framework reduces FA rate effectively without obviously hurting keyword recognition accuracy. Specifically, we can get impressively 0.13 FA per hour on a challenging dataset, with over 90% relative reduction on FA comparing to the transducer based detection model, while keyword recognition accuracy only drops less than 2%.

* Accepted by Interspeech 2022 
Viaarxiv icon

Minimizing Sequential Confusion Error in Speech Command Recognition

Jul 04, 2022
Zhanheng Yang, Hang Lv, Xiong Wang, Ao Zhang, Lei Xie

Figure 1 for Minimizing Sequential Confusion Error in Speech Command Recognition
Figure 2 for Minimizing Sequential Confusion Error in Speech Command Recognition

Speech command recognition (SCR) has been commonly used on resource constrained devices to achieve hands-free user experience. However, in real applications, confusion among commands with similar pronunciations often happens due to the limited capacity of small models deployed on edge devices, which drastically affects the user experience. In this paper, inspired by the advances of discriminative training in speech recognition, we propose a novel minimize sequential confusion error (MSCE) training criterion particularly for SCR, aiming to alleviate the command confusion problem. Specifically, we aim to improve the ability of discriminating the target command from other commands on the basis of MCE discriminative criteria. We define the likelihood of different commands through connectionist temporal classification (CTC). During training, we propose several strategies to use prior knowledge creating a confusing sequence set for similar-sounding command instead of creating the whole non-target command set, which can better save the training resources and effectively reduce command confusion errors. Specifically, we design and compare three different strategies for confusing set construction. By using our proposed method, we can relatively reduce the False Reject Rate~(FRR) by 33.7% at 0.01 False Alarm Rate~(FAR) and confusion errors by 18.28% on our collected speech command set.

* Accepted by Interspeech 2022 
Viaarxiv icon

Decentralized Task Offloading in Edge Computing: A Multi-User Multi-Armed Bandit Approach

Dec 28, 2021
Xiong Wang, Jiancheng Ye, John C. S. Lui

Figure 1 for Decentralized Task Offloading in Edge Computing: A Multi-User Multi-Armed Bandit Approach
Figure 2 for Decentralized Task Offloading in Edge Computing: A Multi-User Multi-Armed Bandit Approach
Figure 3 for Decentralized Task Offloading in Edge Computing: A Multi-User Multi-Armed Bandit Approach
Figure 4 for Decentralized Task Offloading in Edge Computing: A Multi-User Multi-Armed Bandit Approach

Mobile edge computing facilitates users to offload computation tasks to edge servers for meeting their stringent delay requirements. Previous works mainly explore task offloading when system-side information is given (e.g., server processing speed, cellular data rate), or centralized offloading under system uncertainty. But both generally fall short to handle task placement involving many coexisting users in a dynamic and uncertain environment. In this paper, we develop a multi-user offloading framework considering unknown yet stochastic system-side information to enable a decentralized user-initiated service placement. Specifically, we formulate the dynamic task placement as an online multi-user multi-armed bandit process, and propose a decentralized epoch based offloading (DEBO) to optimize user rewards which are subjected under network delay. We show that DEBO can deduce the optimal user-server assignment, thereby achieving a close-to-optimal service performance and tight O(log T) offloading regret. Moreover, we generalize DEBO to various common scenarios such as unknown reward gap, dynamic entering or leaving of clients, and fair reward distribution, while further exploring when users' offloaded tasks require heterogeneous computing resources. Particularly, we accomplish a sub-linear regret for each of these instances. Real measurements based evaluations corroborate the superiority of our offloading schemes over state-of-the-art approaches in optimizing delay-sensitive rewards.

* INFOCOM 2022 
Viaarxiv icon

Efficient Conformer with Prob-Sparse Attention Mechanism for End-to-EndSpeech Recognition

Jun 17, 2021
Xiong Wang, Sining Sun, Lei Xie, Long Ma

Figure 1 for Efficient Conformer with Prob-Sparse Attention Mechanism for End-to-EndSpeech Recognition
Figure 2 for Efficient Conformer with Prob-Sparse Attention Mechanism for End-to-EndSpeech Recognition
Figure 3 for Efficient Conformer with Prob-Sparse Attention Mechanism for End-to-EndSpeech Recognition
Figure 4 for Efficient Conformer with Prob-Sparse Attention Mechanism for End-to-EndSpeech Recognition

End-to-end models are favored in automatic speech recognition (ASR) because of their simplified system structure and superior performance. Among these models, Transformer and Conformer have achieved state-of-the-art recognition accuracy in which self-attention plays a vital role in capturing important global information. However, the time and memory complexity of self-attention increases squarely with the length of the sentence. In this paper, a prob-sparse self-attention mechanism is introduced into Conformer to sparse the computing process of self-attention in order to accelerate inference speed and reduce space consumption. Specifically, we adopt a Kullback-Leibler divergence based sparsity measurement for each query to decide whether we compute the attention function on this query. By using the prob-sparse attention mechanism, we achieve impressively 8% to 45% inference speed-up and 15% to 45% memory usage reduction of the self-attention module of Conformer Transducer while maintaining the same level of error rate.

Viaarxiv icon