Lei He

Library of Congress, Washington, Washington, D.C., United States

Are you Lei He?

Claim your profile

Publications (15)8.17 Total impact

  • [Show abstract] [Hide abstract]
    ABSTRACT: In this paper, we introduce a new matching method based on a novel locally affine-invariant geometric constraint and linear programming techniques. To model and solve the matching problem in a linear programming formulation, all geometric constraints should be able to be exactly or approximately reformulated into a linear form. This is a major difficulty for this kind of matching algorithms. We propose a novel locally affine-invariant constraint which can be exactly linearized and requires a lot fewer auxiliary variables than other linear programming based methods do. The key idea behind it is that each point in the template point set can be exactly represented by an affine combination of its neighboring points, whose weights can be solved easily by least squares. Errors of reconstructing each matched point using such weights are used to penalize the disagreement of geometric relationships between the template points and the matched points. The resulting overall objective function can be solved efficiently by linear programming techniques. Our experimental results on both rigid and non-rigid object matching show the effectiveness of the proposed algorithm.
    IEEE Transactions on Pattern Analysis and Machine Intelligence 04/2012; 35(2):411-24. · 4.80 Impact Factor
  • [Show abstract] [Hide abstract]
    ABSTRACT: This paper presents an overview of the image analysis techniques in the domain of histopathology, specifically, for the objective of automated carcinoma detection and classification. As in other biomedical imaging areas such as radiology, many computer assisted diagnosis (CAD) systems have been implemented to aid histopathologists and clinicians in cancer diagnosis and research, which have been attempted to significantly reduce the labor and subjectivity of traditional manual intervention with histology images. The task of automated histology image analysis is usually not simple due to the unique characteristics of histology imaging, including the variability in image preparation techniques, clinical interpretation protocols, and the complex structures and very large size of the images themselves. In this paper we discuss those characteristics, provide relevant background information about slide preparation and interpretation, and review the application of digital image processing techniques to the field of histology image analysis. In particular, emphasis is given to state-of-the-art image segmentation methods for feature extraction and disease classification. Four major carcinomas of cervix, prostate, breast, and lung are selected to illustrate the functions and capabilities of existing CAD systems.
    Computer methods and programs in biomedicine 03/2012; 107(3):538-56. · 1.56 Impact Factor
  • Source
    [Show abstract] [Hide abstract]
    ABSTRACT: This paper presents a new pixel labeling algorithm for complex histology image segmentation. For each image pixel, a Gaussian mixture model is applied to estimate its neighborhood intensity distributions. With this local distribution fitting, a set of pixels having a full set of source classes (e.g. nuclei, stroma, connective tissue, and background) in their neighborhoods are identified as the seeds for pixel labeling. A seed pixel is labeled by measuring its intensity distance to each of its neighborhood distributions, and the one with the shortest distance is selected to label the seed. For non-seed pixels, we propose two different labeling schemes: global voting and local clustering. In global voting each seed classifies a non-seed pixel into one of the seed's local distributions, i.e., it casts one vote; the final label for the non-seed pixel is the class which gets the most votes, across all the seeds. In local clustering, each non-seed pixel is labeled by one of its own neighborhood distributions. Because the local distributions in a non-seed pixel neighborhood do not necessarily correspond to distinct source classes (i.e., two or more local distributions may be produced by the same source class), we first identify the "true" source class of each local distribution by using the source classes of the seed pixels and a minimum distance criterion to determine the closest source class. The pixel can then be labeled as belonging to this class. With both labeling schemes, experiments on a set of uterine cervix histology images show encouraging performance of our algorithm when compared with traditional multithresholding and K-means clustering, as well as state-of-the-art mean shift clustering, multiphase active contours, and Markov random field-based algorithms.
    Proc SPIE 03/2011;
  • Source
    [Show abstract] [Hide abstract]
    ABSTRACT: Hemiparesis is the most common impairment after stroke, the leading cause of adult disability in the United States. The initial severity of hemiparesis had been the strongest predictor of neuromotor functional recovery level. However, the intervention response of stroke survivors does not always correlate with their initial level of impairment. This implies the existence of other factors that may significantly affect stroke survivors' recovery process. In order to design targeting intervention therapy strategies, it is critical to consider these factors in a principled, comprehensive way so that physical rehabilitation (PR) researchers may predict which stroke survivors will respond best to therapy and subsequently, determine if a particular type of therapy is a more optimal match. Currently, such prediction is primarily a manual process and remains a challenging task to PR researchers and clinicians. We propose a computing framework based upon a domain-specific ontology. This framework aims to facilitate knowledge acquisition from existing sources via semantics-enhanced data mining (SEDM) techniques. As a result, it will assist PR researchers and clinicians in better predicting stroke survivors' neuromotor functional recovery level, and will help physical therapists customize most effective intervention therapy plans for individual stroke survivors.
    Bioinformatics and Biomedicine Workshops (BIBMW), 2010 IEEE International Conference on; 01/2011
  • Source
    [Show abstract] [Hide abstract]
    ABSTRACT: Hemiparesis is the most common impairment after stroke, and the initial severity of hemiparesis had been the strongest predictor of neuromotor functional recovery level. However, the intervention response of stroke survivors does not always correlate with their initial level of impairment, which implies the existence of other factors that may significantly affect stroke survivors' recovery process. It is critical to consider these factors in a principled, comprehensive way so that physical rehabilitation (PR) researchers may predict which stroke survivors will respond best to therapy and, as a result, to determine if a particular type of therapy is a more optimal match. Currently, such prediction is primarily a manual process and remains a challenging task to PR researchers and clinicians. Based upon a domain-specific ontology, NeuMORE, we propose a computing framework that aims to facilitate knowledge acquisition from existing sources via semantics-enhanced data mining (SEDM) techniques. It will assist PR researchers and clinicians in better predicting stroke survivors' neuromotor functional recovery level, and will help physical therapists customize most effective intervention therapy plans for individual stroke survivors.
    Bioinformatics and Biomedicine Workshops (BIBMW), 2010 IEEE International Conference on; 01/2011
  • [Show abstract] [Hide abstract]
    ABSTRACT: Effective assessment is vital in educational activities. We propose IWAS (intelligent Web-based assessment system), an intelligent, generalized and real-time system to assess both learning and teaching. IWAS provides a foundation for more efficiency in instructional activities and, ultimately, students' performances. Our contributions are summarized as: (1) Given the causes (student knowledge levels and learning styles), BN (Bayesian Networks) technique is utilized to automatically reason on the probabilities of the presence of the effects (learning outcomes); (2) The absence of teaching practice assessments is addressed via the feedbacks from three different levels, aiming to correlate the teaching assessment with the learning assessments for the improved effectiveness in instructional activities; (3) Under a client/server architecture, IWAS is decomposed into a set of independent modules; through the standard inter-module interfaces, the flexibility of easy maintenance makes IWAS a generalized system adaptable to different domains; and (4) Web technologies are integrated to deliver the formative feedbacks to users in a timely manner. (Contains 1 table and 2 figures.)
    Online Submission. 01/2011;
  • [Show abstract] [Hide abstract]
    ABSTRACT: This paper presents a new object tracking model that systematically combines region and boundary features. Besides traditional region features (intensity/color and texture), we design a new boundary-based object detector for accurate and robust tracking in low-contrast and complex scenes, which usually appear in the commonly used monochrome surveillance systems. In our model, region feature-based energy terms are characterized by probability models, and boundary feature terms include edge and frame difference. With a new weighting term, a novel energy functional is proposed to systematically combine the region and boundary-based components, and it is minimized by a level set evolution equation. For an efficient computational cost, motion information is utilized for new frame level set initialization. Compared with region feature-based models, the experimental results show that the proposed model significantly improves the performance under different circumstances, especially for objects in low-contrast and complex environments.
    IEEE Transactions on Circuits and Systems for Video Technology 01/2011; 21:1784-1794. · 1.82 Impact Factor
  • [Show abstract] [Hide abstract]
    ABSTRACT: In this paper we present a multiphase level set model for histology image segmentation. Global K-means energy is weighted by a Gaussian kernel to cluster image pixels in local neighborhoods. We group these local clusters into different source classes using a multiphase level set model to produce the final segmentation results. Our energy functional is formulated as the integral of local K-means energies across the entire image. Unlike current local region-based active contour methods that update the pixel neighborhood distributions (e.g. local intensity means) in each iteration, we estimate these statistics before contour evolution for more efficient computation. In addition, such pre-derived local intensity distributions enable a model without initial contour selection, i.e., the level set functions can be initialized with a random constant instead of a distance map. In this way our model ameliorates the initialization sensitivity problem of most active contour methods. Experiments on the National Cancer Institute ALTS histology images show the improved performance of our approach over standard multithresholding and K-means clustering, as well as state-of-the-art active contours, mean shift clustering, and Markov random field-based pixel labeling methods.
    2011 IEEE International Conference on Healthcare Informatics, Imaging and Systems Biology, HISB 2011, San Jose, CA, USA, July 26-29, 2011; 01/2011
  • Source
    [Show abstract] [Hide abstract]
    ABSTRACT: The identification and characterization of important roles microRNAs (miRNAs) played in human cancer is an increasingly active area in medical informatics. In particular, the prediction of miRNA target genes remains a challenging task to cancer researchers. Current efforts have focused on manual knowledge acquisition from existing miRNA databases, which is time-consuming, error-prone, and subject to biologists’ limited prior knowledge. Therefore, an effective knowledge acquisition has been inhibited. We propose a computing framework based on the Ontology for MicroRNA Target Prediction (OMIT), the very first ontology in miRNA domain. With such formal knowledge representation, it is thus possible to facilitate knowledge discovery and sharing from existing sources. Consequently, the framework aims to assist biologists in unraveling important roles of miRNAs in human cancer, and thus to help clinicians in making sound decisions when treating cancer patients.
    10/2010: pages 1160-1167;
  • Source
    [Show abstract] [Hide abstract]
    ABSTRACT: Worldwide health scientists are producing, accessing, analyzing, integrating, and storing massive amounts of digital medical data daily, through observation, experimentation, and simulation. If we were able to effectively transfer and integrate data from all possible resources, then a deeper understanding of all these data sets and better exposed knowledge, along with appropriate insights and actions, would be granted. Unfortunately, in many cases, the data users are not the data producers, and they thus face challenges in harnessing data in unforeseen and unplanned ways. In order to obtain the ability to integrate heterogeneous data, and thereby efficiently revolutionize the traditional medical and biological research, new methodologies built upon the increasingly pervasive cyberinfrastructure are required to conceptualize traditional medical and biological data, and acquire the “deep” knowledge out of original data thereafter. As formal knowledge representation models, ontologies can render invaluable help in this regard. In this paper, we summarize the state-of-the-art research in ontological techniques and their innovative application in medical and biological areas.
    Fuzzy Systems and Knowledge Discovery (FSKD), 2010 Seventh International Conference on; 09/2010
  • Source
    [Show abstract] [Hide abstract]
    ABSTRACT: In this paper, we present a new object matching algorithm based on linear programming and a novel locally affine-invariant geometric constraint. Previous works have shown possible ways to solve the feature and object matching problem by linear programming techniques. To model and solve the matching problem in a linear formulation, all geometric constraints should be able to be exactly or approximately reformulated into a linear form. This is a major difficulty for this kind of matching algorithms. We propose a novel locally affine-invariant constraint which can be exactly linearized and requires a lot fewer auxiliary variables than the previous work does. The key idea behind it is that each point can be exactly represented by an affine combination of its neighboring points, whose weights can be solved easily by least squares. The resulting overall objective function can then be solved efficiently by linear programming techniques. Our experimental results on both rigid and non-rigid object matching show the advantages of the proposed algorithm.
    Computer Vision and Pattern Recognition (CVPR), 2010 IEEE Conference on; 07/2010
  • Source
    [Show abstract] [Hide abstract]
    ABSTRACT: The identification and characterization of important roles microRNAs (miRNAs) played in human cancer is an increasingly active area in medical informatics, and the prediction of miRNA target genes remains a challenging task to cancer researchers. We propose an innovative computing framework based on the Ontology for MicroRNA Target (OMIT) to facilitate knowledge acquisition from existing sources. The project aims to assist biologists in unraveling important roles of miRNAs in human cancer, and thus to help clinicians in making sound decisions when treating cancer patients.
    Proceedings of the First ACM International Conference on Bioinformatics and Computational Biology, BCB 2010, Niagara Falls, NY, USA, August 2-4, 2010; 01/2010
  • [Show abstract] [Hide abstract]
    ABSTRACT: This paper presents a new stereo vision-based model for multi-object detection and tracking in surveillance systems. Unlike most existing monocular camera-based systems, a stereo vision system is constructed in our model to overcome the problems of illumination variation, shadow interference, and object occlusion. In each frame, a sparse set of feature points are identified in the camera coordinate system, and then projected to the 2D ground plane. A kernel-based clustering algorithm is proposed to group the projected points according to their height values and locations on the plane. By producing clusters, the number, position, and orientation of objects in the surveillance scene can be determined for online multi-object detection and tracking. Experiments on both indoor and outdoor applications with complex scenes show the advantages of the proposed system.
    Pattern Recognition. 01/2010;
  • Source
    [Show abstract] [Hide abstract]
    ABSTRACT: This paper presents a new algorithm for hematoxylin and eosin (H&E) stained histology image segmentation. With both local and global clustering, Gaussian mixture models (GMMs) are applied sequentially to extract tissue constituents such as nuclei, stroma, and connecting contents from background. Specifically, local GMM is firstly applied to detect nuclei by scanning the input image, which is followed by global GMM to separate other tissue constituents from background. Regular RGB (red, green and blue) color space is employed individually for the local and global GMMs to make use of the H&E staining features. Experiments on a set of cervix histology images show the improved performance of the proposed algorithm when compared with traditional K-means clustering and state-of-art multiphase level set methods.
    10th International Conference on Hybrid Intelligent Systems (HIS 2010), Atlanta, GA, USA, August 23-25, 2010; 01/2010
  • [Show abstract] [Hide abstract]
    ABSTRACT: Image verification has been widely used in numerous websites to prevent them from batch registration or automated posting. One category of the image verification is generated by adding noise into character or digit images to make them hard to be recognized by Optical Character Recognition (OCR). In this paper, we propose a novel probability gradient function for active contour models to efficiently segment this type of images for easier recognition. Experiments on a set of images with different intensities and types of noise show the superiority of the proposed probability gradient to traditional method. The purpose of our paper is to warn some websites who are still using such kind of verification: they should improve their defense method to prevent them from the potential risk.
    Computer Vision - ACCV 2009, 9th Asian Conference on Computer Vision, Xi'an, China, September 23-27, 2009, Revised Selected Papers, Part III; 01/2009

Publication Stats

42 Citations
8.17 Total Impact Points

Institutions

  • 2012
    • Library of Congress
      Washington, Washington, D.C., United States
  • 2010–2012
    • Lehigh University
      • Department of Computer Science and Engineering
      Lehigh Acres, FL, United States
    • National Institutes of Health
      • Division of Cancer Epidemiology and Genetics
      Maryland, United States
    • Shanghai Jiao Tong University
      Shanghai, Shanghai Shi, China
    • Armstrong State University
      • College of Science & Technology
      Savannah, Georgia, United States