Max Q.-H. Meng

The Chinese University of Hong Kong, Hong Kong, Hong Kong

Are you Max Q.-H. Meng?

Claim your profile

Publications (372)129.71 Total impact

  • Haiying Liu · Jason Gu · Max Q.-H. Meng · Wu-Sheng Lu

    No preview · Article · Jan 2016
  • Yixuan Yuan · Max Q. -H. Meng
    [Show abstract] [Hide abstract]
    ABSTRACT: Wireless capsule endoscopy (WCE) is a revolutionary imaging technique that enables direct inspection of the gastrointestinal tract in a non-invasive way. However, viewing the large amounts of images is a very time-consuming and labor intensive task for clinicians. In this paper, we propose an automatic bleeding detection method in the WCE images. We propose a two-stage saliency map extraction method to highlight bleeding regions where the first-stage saliency map is created by means of different color channels mixer and the second-stage saliency map is obtained from the visual contrast in the RGB color space. Followed by an appropriate fusion strategy and threshold, we localize the bleeding areas in the WCE images. Then we extract statistic color features in the corresponding saliency region and non-saliency region respectively and fuse them together to represent the whole WCE images. Finally Support Vector Machine (SVM) is applied to carry out the experiment on 800 sample WCE images. Experiment result achieves an accuracy of 95.89%, sensitivity of 98.77% and specificity of 93.45%. This inspiring result demonstrates that the proposed method is very effective in detecting bleeding patterns in the WCE images. Our comparison studies with several state-of-the-art bleeding detection methods confirm that the proposed method achieves much better results than those of the alternative techniques.
    No preview · Article · Jun 2015 · Proceedings - IEEE International Conference on Robotics and Automation
  • Source
    Minhua Zheng · · Ajung Moon · Elizabeth A Croft · Max Q.-H. Meng
    [Show abstract] [Hide abstract]
    ABSTRACT: In this paper, we investigate the use of a robot's gaze to improve the timing and subjective experience of face-to-face robot-to-human handovers. Based on observations of human gaze behaviors during face-to-face human–human handovers, we implement various gaze behaviors on a PR2 humanoid robot. We conducted two consecutive robot-to-human handover studies. Results show that when the robot continually gazes at a projected handover position while handing over an object, the human receivers reach for the object significantly earlier than when the robot looks down, away from the handover location; further, when the robot continually gazes at the receiver's face instead of the handover position, the receivers reach for the object even earlier. When the robot—instead of continually gazing at a location—transitions its gaze from the handover position to the receivers' face, or vice versa, the receivers' reach time did not improve; however, the receivers perceive these gaze transitions to better communicate handover timing than continual gazes. Finally, the receivers perceive the robot to be more likeable and anthropomorphic when it looks at their face than when it does not. Findings from our studies indicate that robot's use of gaze can help improve both fluency and subjective experience of the robot-to-human handover interactions.
    Full-text · Article · Jun 2015 · International Journal of Social Robotics
  • Wenzheng Chi · Jiaole Wang · Max Q.-H. Meng
    [Show abstract] [Hide abstract]
    ABSTRACT: The state-of-the-art person identification and verification technologies require apparent human biological features as inputs explicitly. In many special scenarios, however, it is impossible to obtain the apparent human biological features since such features usually require intrusive human interaction directly. In this paper, we have proposed a hierarchical skeleton feature model (HSFM) for person verification by using implicit biometric skeletal data estimated from the emerging RGB-D camera. To validate the viability of the proposed feature model, experiments have been carried out to retrieve the estimated skeletal data from 11 volunteers by using the RGB-D camera. 2fold cross-validations with SVM (Support Vector Machine) and KNN (K-Nearest Neighbors) methods were adopted and carried out 200 times to train and test the obtained data for regarding each participant as the authorized person. Furthermore, the parameter accuracy, sensitivity, specificity and time cost of the proposed human verification method were analyzed in the experiments. The intensive analyses on our data set have shown the feasibility to use the estimated human skeletal data from RGB-D camera for human verification.
    No preview · Article · Apr 2015
  • Yixuan Yuan · Jiaole Wang · Baopu Li · Max Q. -H. Meng
    [Show abstract] [Hide abstract]
    ABSTRACT: Ulcer is one of the most common symptoms of many serious diseases in the human digestive tract. Especially for the ulcers in the small bowel where other procedures cannot adequately visualize, wireless capsule endoscopy (WCE) is increasingly being used in the diagnosis and clinical management. Because WCE generates large amount of images from the whole process of inspection, computer-aided detection of ulcer is considered an indispensable relief to clinicians. In this paper, a two-staged fully automated computer-aided detection system is proposed to detect ulcer from WCE images. In the first stage, we propose an effective saliency detection method based on multi-level superpixel representation to outline the ulcer candidates. To find the perceptually and semantically meaningful salient regions, we first segment the image into multi-level superpixel segmentations. Each level corresponds to different initial region sizes of the superpixels. Then we evaluate the corresponding saliency according to the color and texture features in superpixel region of each level. In the end, we fuse the saliency maps from all levels together to obtain the final saliency map. In the second stage, we apply the obtained saliency map to better encode the image features for the ulcer image recognition tasks. Because the ulcer mainly corresponds to the saliency region, we propose a saliency max-pooling method integrated with the Locality-constrained Linear Coding (LLC) method to characterize the images. Experiment results achieve promising 92.65% accuracy and 94.12% sensitivity, validating the effectiveness of the proposed method. Moreover, the comparison results show that our detection system outperforms the state-ofthe- art methods on the ulcer classification task.
    No preview · Article · Apr 2015
  • Baopu Li · Can Yang · Tianfu Wang · Guoqing Xu · Qi Zhang · Max Q.-H. Meng · Chao Hu
    [Show abstract] [Hide abstract]
    ABSTRACT: Capsule endoscopy is a new imaging technology for small intestine due to its breakthrough for direct visualization of small intestine for the first time. However, the video data produced for each patient costs a physician much time to inspect. Aiming for reducing the burden of a physician, video scene analysis is indispensable. In this paper, we propose a new video segmentation method to analyze a CE video data since video segmentation is the first step in video scene analysis. A novel color textural feature is utilized to describe the content of the frame in a CE video, then spectral clustering method is applied to segment a CE video into meaningful parts via shot boundary detection. Preliminary experiments on ten short CE videos demonstrate a promising performance of the proposed scheme.
    No preview · Article · Mar 2015
  • Qi Zhang · Baopu Li · G.-Q. Xu · Yimin Zhou · Ming Wang · Max Q.-H Meng
    [Show abstract] [Hide abstract]
    ABSTRACT: The second generation of Kinect sensor is a revolutionary multiple cameras including RGB camera, infrared camera to differentiate depth, infrared camera and audio array. In this paper, we first introduce the differences between the Kinect and Kinect2.0 sensor. Then a system of Kinect2.0 based human interaction system for mobile robot is presented. The function of skeleton tracking of Kinect2.0 is used to identify a specific person in the depth image. Finally, people's heart rate detection method is proposed using RGB camera of Kinect2.0. Preliminary experiments verify that the designed measurement system is effective.
    No preview · Article · Mar 2015
  • Source
    [Show abstract] [Hide abstract]
    ABSTRACT: A robot that can fluently hand over objects to people can be useful in many applications. In an effort to develop a fluent robot-to-human handover system, this work investigates people's behavioural responses to a robot that hands over objects to them while using different types of gaze cues. In our previous work, we found empirical evidence that the use of a robot's head gaze can affect a person's timing of reaching towards the offered object. In this paper, we investigate this effect further by exploring the manner in which human's reaching and gaze behaviours are affected by a robot's head gaze. We conducted a video-based investigation of 97 naïve participants' behavioural responses to robot-to-human handovers. Through a frame-by-frame analysis, we recorded a detailed timeline of the robot's and human's gaze and reaching behaviours. Results confirm the finding from our previous study that the robot's head gaze can significantly impact the timing of human receiver's reaching behaviour during handovers. In addition, our results demonstrate that the robot's head gaze affects human's gaze behaviour during handovers, and this effect explains some unexpected findings in our previous work.
    Full-text · Conference Paper · Dec 2014
  • Kun Li · Max Meng
    [Show abstract] [Hide abstract]
    ABSTRACT: Most robot semantic mapping methods only consider the intrinsic properties of landmarks and objects inside a scene, by detecting them with their appearances, and some other methods include extrinsic properties with manually designed object relations. In this work, we use relational operators to capture the extrinsic property values, and adopt conditional random field to integrate intrinsic and extrinsic property values into semantic mapping. We compare our approach with three types of semantic maps, and show that our approach allows the robot to find designated objects more accurately.
    No preview · Conference Paper · Dec 2014
  • Yuxiang Sun · Ming Liu · Max Q.-H Meng
    [Show abstract] [Hide abstract]
    ABSTRACT: Due to the unavailable GPS signals in indoor environments, indoor localization has become an increasingly heated research topic in recent years. Researchers in robotics community have tried many approaches, but this is still an unsolved problem considering the balance between performance and cost. The widely deployed low-cost WiFi infrastructure provides a great opportunity for indoor localization. In this paper, we develop a system for WiFi signal strength-based indoor localization and implement two approaches. The first is improved KNN algorithm-based fingerprint matching method, and the other is the Gaussian Process Regression (GPR) with Bayes Filter approach. We conduct experiments to compare the improved KNN algorithm with the classical KNN algorithm and evaluate the localization performance of the GPR with Bayes Filter approach. The experiment results show that the improved KNN algorithm can bring enhancement for the fingerprint matching method compared with the classical KNN algorithm. In addition, the GPR with Bayes Filter approach can provide about 2m localization accuracy for our test environment.
    No preview · Article · Oct 2014
  • Source
    Kun Li · Max Q. -H. Meng
    [Show abstract] [Hide abstract]
    ABSTRACT: To learn object models for robotic manipulation, unsupervised methods cannot provide accurate object structural information and supervised methods require a large amount of manually labeled training samples, thus interactive object segmentation is developed to automate object modeling. In this article, we formulate a novel dynamic process for interactive object segmentation, and develop a solution based on particle filter and active learning so that a robot can manipulate and learn object structures incrementally and automatically. We demonstrate our method with a humanoidrobot on different types of objects, and compare its segmentation performancewith established methods on selected objects. The result shows that our approach allows more accurate object modeling and reveals richer object structural information.
    Preview · Article · Aug 2014
  • Source
    Kun Li · Max Q. -H. Meng
    [Show abstract] [Hide abstract]
    ABSTRACT: In a robot-centered smart home, the robot observes the home states with its own sensors, and then it can change certain object states according to an operator's commands for remote operations, or imitate the operator's behaviors in the house for autonomous operations. To model the robot's imitation of the operator's behaviors in a dynamic indoor environment, we use multi-relational chains to describe the changes of environment states, and apply inverse reinforcement learning to encoding the operator's behaviors with a learned reward function. We implement this approach with a mobile robot, and do five experiments to include increasing training days, object numbers, and action types. Besides, a baseline method by directly recording the operator's behaviors is also implemented, and comparison is made on the accuracy of home state evaluation and the accuracy of robot action selection. The results show that the proposed approach handles dynamic environment well, and guides the robot's actions in the house more accurately.
    Preview · Article · Aug 2014
  • Jiaole Wang · Hongliang Ren · Max Meng

    No preview · Conference Paper · Jun 2014
  • Kun Li · Max Meng
    [Show abstract] [Hide abstract]
    ABSTRACT: The performance of robotic object manipulation relies heavily on the selection of object model. In this article, we develop a multilevel part-based object model by applying latent support vector machine to training a hierarchical object structure. We implement our method with a robot arm and a depth sensor in Robot Operating System, and then we compare the recognition performance of this model with established methods on a point cloud data set and show the manipulation performance of our model on three practical tasks. The result demonstrates that our robot recognizes and manipulates objects more accurately with this multilevel part-based object model.
    No preview · Conference Paper · May 2014
  • Source
    Lujia Wang · Ming Liu · Max Q.-H Meng
    [Show abstract] [Hide abstract]
    ABSTRACT: In order to share information in the cloud for multi-robot systems, efficient data transmission is essential for real-time operations such as coordinated robotic missions. As a limited resource, bandwidth is ubiquitously required by applications among physical multi-robot systems. In this paper, we proposed a hierarchical auction-based mechanism, namely LQM (Link Quality Matrix)-auction. It consists of multiple procedures, such as hierarchical auction, proxy scheduling. Note that the proposed method is designed for real-time resource retrieval for physical multi-robot systems, instead of simulated virtual agents. We validate the proposed mechanism through real-time experiments. The results show that LQM-auction is suitable for scheduling a group of robots, leading to optimized performance for resource retrieval.
    Full-text · Conference Paper · May 2014
  • Yangming Li · Shuai Li · Quanjun Song · Hai Liu · Max Q.-H. Meng
    [Show abstract] [Hide abstract]
    ABSTRACT: Data association is a fundamental problem in multisensor fusion, tracking, and localization. The joint compatibility test is commonly regarded as the true solution to the problem. However, traditional joint compatibility tests are computationally expensive, are sensitive to linearization errors, and require the knowledge of the full covariance matrix of state variables. The paper proposes a posterior-based joint compatibility test scheme to conquer the three problems mentioned above. The posterior-based test naturally separates the test of state variables from the test of observations. Therefore, through the introduction of the robot movement and proper approximation, the joint test process is sequentialized to the sum of individual tests; therefore, the test has O(n) complexity (compared with O(n2) for traditional tests), where n denotes the total number of related observations. At the same time, the sequentialized test neither requires the knowledge to the full covariance matrix of state variables nor is sensitive to linearization errors caused by poor pose estimates. The paper also shows how to apply the proposed method to various simultaneous localization and mapping (SLAM) algorithms. Theoretical analysis and experiments on both simulated data and popular datasets show the proposed method outperforms some classical algorithms, including sequential compatibility nearest neighbor (SCNN), random sample consensus (RANSAC), and joint compatibility branch and bound (JCBB), on precision, efficiency, and robustness.
    No preview · Article · Feb 2014 · IEEE Transactions on Industrial Informatics
  • Jiaole Wang · Hongliang Ren · Max Meng

    No preview · Article · Jan 2014 · IEEE Transactions on Automation Science and Engineering
  • Chao Hu · Wentai Qu · Zhihuan Zhang · Zhongqing Feng · Shuang Song · Max Q.-H. Meng
    [Show abstract] [Hide abstract]
    ABSTRACT: In this paper, a novel method is proposed to recover and extract the original signal parameters from the saturated multifrequency sinusoid wave signals. It makes use of the zero-crossing characteristics of the multifrequency sinusoid signals, to collect valid samples in unsaturated parts of the signals. On these valid samples, the amplitudes and phases of the specific original ac sensing signals can be linearly computed by applying the least square method. The simulation results show that the proposed method has satisfactory accuracy even with very large saturation ( ~ 10 times of the saturation limit) and large dc offset, which frees us from the restriction to avoid the signal saturation problem in the signal acquisition. The method is realized by the software algorithm, and no longer requires the common used hardware-the phase sensitive detection circuit. Hence when it is applied to the magnetic coupling system, we will obtain much simpler system composition, higher accuracy, and high execution speed.
    No preview · Article · Nov 2013 · IEEE Sensors Journal
  • Source
    Lujia Wang · Ming Liu · Max Q.-H. Meng
    [Show abstract] [Hide abstract]
    ABSTRACT: Networked multi-robot systems benefit from a large amount of heterogeneous online data on the server, and enable poor-equipped robots to fulfill complex tasks. However, as a major bottleneck of practical network, the limited bandwidth is lack of consideration. In the matter of fact, resource competition is pervasive for practical networked robotic applications. We propose a multi-robot negotiation mechanism in this paper. It includes a game theory based auction for allocating resources that are shared among robot clients, such as the network bandwidth. We validate the proposed strategy by a joint-surveillance scenario. Experimental results demonstrate that the proposed framework achieves excellent Quality of Service (QoS) performance under the condition of resource competition, where a shared network with limited bandwidth is optimized.
    Full-text · Conference Paper · Aug 2013
  • Yixuan Yuan · Max Q.-H. Meng
    [Show abstract] [Hide abstract]
    ABSTRACT: Wireless capsule endoscopy (WCE) is an advanced, patient-friendly imaging technique that enables close examination of the entire small intestine. Since it usually takes hours to review all the video data even by professional clinician, the automatic computer-aided technique is highly demanded. This paper presents a hierarchical methodology for detecting key frames in WCE images. In the first stage, we choose key frames whose changes of information entropies take the local maximum by automatic threshold to cut the images into several sub clots. Then AP clustering method is applied in each clot to extract the second stage key frames. Our method maintains the temporal information and maximizes the content distance. Experimental results demonstrate that the proposed techniques achieve inspiring performance with fidelity 0.9206 and compression ratio 0.9125 on average.
    No preview · Conference Paper · Aug 2013

Publication Stats

3k Citations
129.71 Total Impact Points

Institutions

  • 2003-2015
    • The Chinese University of Hong Kong
      • Department of Electronic Engineering
      Hong Kong, Hong Kong
  • 2013
    • Hong Kong SAR Government
      Hong Kong, Hong Kong
  • 2004-2011
    • Chinese Academy of Sciences
      • • Shenzhen Institutes of Advanced Technology
      • • Institute of Intelligent Machines
      Peping, Beijing, China
    • University of Ottawa
      Ottawa, Ontario, Canada
  • 2010
    • National Space Science
      Peping, Beijing, China
    • University of Science and Technology of China
      Luchow, Anhui Sheng, China
  • 2009
    • Johns Hopkins University
      • Department of Biomedical Engineering
      Baltimore, MD, United States
    • Dalian University of Technology
      Lü-ta-shih, Liaoning, China
  • 2008
    • Ningbo University
      Ning-po, Zhejiang Sheng, China
  • 2007
    • Hefei University of Technology
      Luchow, Anhui Sheng, China
    • China Three Gorges University
      Tung-hu, Hubei, China
    • Carleton University
      • Department of Systems and Computer Engineering
      Ottawa, Ontario, Canada
  • 2006
    • Shandong University
      • School of Control Science and Engineering
      Chi-nan-shih, Shandong Sheng, China
    • Chongqing University of Posts and Telecommunications
      Ch’ung-ch’ing-shih, Chongqing Shi, China
  • 2005
    • Hefei Institute of Physical Sciences, Chinese Academy of Sciences
      Luchow, Anhui Sheng, China
  • 1999-2004
    • University of Guelph
      • School of Engineering
      Guelph, Ontario, Canada
  • 1995-2004
    • University of Alberta
      • Department of Electrical and Computer Engineering
      Edmonton, Alberta, Canada
  • 2002
    • Harbin Engineering University
      • College of Automation
      Charbin, Heilongjiang Sheng, China
  • 2001
    • Dalhousie University
      • Department of Electrical & Computer Engineering
      Halifax, Nova Scotia, Canada