Browsing by Author "Du, Eliza Y."

Now showing 1 - 5 of 5
  • Thumbnail Image
    Item
    Building a Surface Atlas of Hippocampal Subfields from MRI Scans using FreeSurfer, FIRST and SPHARM
    (Institute of Electrical and Electronics Engineers, 2014-08) Cong, Shan; Rizkalla, Maher; Du, Eliza Y.; West, John; Risacher, Shannon; Saykin, Andrew J.; Shen, Li; Alzheimer's Disease Neuroimaging Initiative; Department of Medicine, IU School of Medicine
    The hippocampus is widely studied with neuroimaging techniques given its importance in learning and memory and its potential as a biomarker for brain disorders such as Alzheimer's disease and epilepsy. However, its complex folding anatomy often presents analytical challenges. In particular, the critical hippocampal subfield information is usually ignored by hippocampal registration in detailed morphometric studies. Such an approach is thus inadequate to accurately characterize hippocampal morphometry and effectively identify hippocampal structural changes related to different conditions. To bridge this gap, we present our initial effort towards building a computational framework for subfield-guided hippocampal morphometry. This initial effort is focused on surface-based morphometry and aims to build a surface atlas of hippocampal subfields. Using the FreeSurfer software package, we obtain valuable hippocampal subfield information. Using the FIRST software package, we extract reliable hippocampal surface information. Using SPHARM, we develop an approach to create an atlas by mapping interpolated subfield information onto an average surface. The empirical result using ADNI data demonstrates the promise and good reproducibility of the proposed method.
  • Thumbnail Image
    Item
    An Extreme Learning Machine-based Pedestrian Detection Method
    (Office of the Vice Chancellor for Research, 2013-04-05) Yang, Kai; Du, Eliza Y.; Delp, Edward J.; Jiang, Pingge; Jiang, Feng; Chen, Yaobin; Sherony, Rini; Takahashi, Hiroyuki
    Pedestrian detection is a challenging task due to the high variance of pedestrians and fast changing background, especially for a single in-car camera system. Traditional HOG+SVM methods have two challenges: (1) false positives and (2) processing speed. In this paper, a new pedestrian detection method using multimodal HOG for pedestrian feature extraction and kernel based Extreme Learning Machine (ELM) for classification is presented. The experimental results using our naturalistic driving dataset show that the proposed method outperforms the traditional HOG+SVM method in both recognition accuracy and processing speed.
  • Thumbnail Image
    Item
    IUPUI Imaging Research Council
    (Office of the Vice Chancellor for Research, 2012-04-13) Hutchins, Gary D.; Wilson, Kathryn J.; Sturek, Michael S.; Du, Eliza Y.; Fletcher, James W.; Long, Eric C.; Molitoris, Bruce A.; Johnson, Daniel P.; Day, Richard N.; Barnett, William K.; Palakal, Mathew J.
    Abstract The IUPUI Imaging Research Council was created by the IUPUI Vice Chancellor for Research to provide guidance and direction for expansion of research imaging initiatives across all Schools and Departments within IUPUI. The specific goals of the council are: • To encourage and coordinate collaboration among IUPUI researchers from different disciplines • To provide advice and guidance in the realization of highly competitive large grant proposals that will support and grow the IUPUI imaging efforts into major nationally and internationally recognized programs • To develop a strategic plan that will enable IUPUI to become nationally and internationally known as the place for imaging research and its applications • To determine strategic areas of strength and growth • To determine available and needed resources • To determine strategic external partnerships Activities organized by the council to date include sponsoring an IUPUI Imaging Research Workshop on November 17, 2011. This workshop consisted of invited presentations, a poster session, and working group breakout sessions. Working groups explored research opportunities and needs in four priority areas (neuroscience, cancer, cardiovascular disease, and remote sensing). The council has recently initiated a monthly seminar series and is actively developing an IUPUI research imaging strategic plan. For more information visit the IUPUI Imaging Research Initiative website at www.imaging.iupui.edu.
  • Thumbnail Image
    Item
    IUPUI Imaging Research Initiative
    (Office of the Vice Chancellor for Research, 2013-04-05) Holland, Mark; Barnett, William; Burr, David B.; Day, Richard; Du, Eliza Y.; Gattone, Vincent, III; Fletcher, James; Johnson, Daniel P.; Long, Eric; Molitoris, Bruce A.; Palakal, Mathew; Salama, Paul; Sturek, Michael; Hutchins, Gary D.
    Imaging has become an essential research tool in a majority of scientific disciplines. The IUPUI Imaging Research Initiative (IRI) has been established to bring together researcher investigators who develop novel imaging technologies with those who utilize imaging tools to advance their research with the primary objective of building a large scale imaging research infrastructure at IUPUI. An Imaging Research Council has been created to establish priorities for the IRI and help guide the development of an IUPUI research imaging infrastructure and sustainable research funding base. The specific goals of the council include: • To encourage and coordinate collaboration among IUPUI researchers from different disciplines • To provide advice and guidance in the realization of highly competitive large grant proposals that will support and grow the IUPUI imaging efforts into major nationally and internationally recognized programs • To develop a strategic plan that will enable IUPUI to become nationally and internationally known as the place for imaging research and its applications • To determine strategic areas of strength and growth • To determine available and needed resources • To determine strategic external partnerships
  • Thumbnail Image
    Item
    Revocable, Interoperable and User-Centric (Active) Authentication Across Cyberspace
    (Office of the Vice Chancellor for Research, 2014-04-11) Sui, Yan; Zou, Xukai; Du, Eliza Y.; Li, Feng
    This work addresses fundamental and challenging user authentication and universal identity issues and solves the problems of system usability, authentication data security, user privacy, irrevocability, interoperability, cross-matching attacks, and post-login authentication breaches associated with existing authentication systems. It developed a solid user-centric biometrics based authentication model, called Bio-Capsule (BC), and implemented an (active) authentication system. BC is the template derived from the (secure) fusion of a user’s biometrics and that of a Reference Subject (RS). RS is simply a physical object such as a doll or an artificial one, such as an image. It is users’ BCs, rather than original biometric templates, that are utilized for user authentication and identification. The implemented (active) authentication system will facilitate and safely protect individuals’ diffused cyber activities, which is particularly important nowadays, when people are immersed in cyberspace. User authentication is the first guard of any trustworthy computing system. Along with people’s immersion in the penetrated cyber space integrated with information, networked systems, applications and mobility, universal identity security& management and active authentication become of paramount importance for cyber security and user privacy. Each of three typical existing authentication methods, what you KNOW (Password/PIN), HAVE (SmartCard), and ARE (Fingerprint/Face/Iris) and their combinations, suffer from their own inherent problems. For example, biometrics is becoming a promising authentication/identification method because it binds an individual with his identity, is resistant to losses, and does not need to memorize/carry. However, biometrics introduces its own challenges. One serious problem with biometrics is that biometric templates are hard to be replaced once compromised. In addition, biometrics may disclose user’s sensitive information (such as race, gender, even health condition), thus creating user privacy concerns. In the recent years, there has been intensive research addressing biometric template security and replaceability, such as cancelable biometrics and Biometric Cryptosystems. Unfortunately, these approaches do not fully exploit biometric advantages (e.g., requiring a PIN), reduce authentication accuracy, and/or suffer from possible attacks. The proposed approach is the first elegant solution to effectively address irreplaceability, privacy-preserving, and interoperability of both login and after-login authentication. Our methodology preserves biometrics’ robustness and accuracy, without sacrificing system acceptability for the same user, and distinguishability between different users. Biometric features cannot be recovered from the user’s Biometric Capsule or Reference Subject, even when both are stolen. The proposed model can be applied at the signal, feature, or template levels, and facilitates integration with new biometric identification methods to further enhance authentication performance. Moreover, the proposed active, non-intrusive authentication is not only scalable, but also particularly suitable to emerging portable, mobile computing devices. In summary, the proposed approach is (i) usercentric, i.e., highly user friendly without additional burden on users, (ii) provably secure and resistant to attacks including cross-matching attacks, (iii) identity-bearing and privacy-preserving, (iv) replaceable, once Biometric Capsule is compromised, (v) scalable and highly adaptable, (vi) interoperable and single signing on across systems, and (vii) cost-effective and easy to use.