XYXiaojing Yuan

Xiaojing Yuan

*This is an algorithmically generated author record based on publicly available information.

Current Organizations / Institutes

The organizations/institutes from the scholar's latest article

Previous Organizations / Institutes

The organizations / institutes from the scholar's other articles in the past ten years


Xiaojing Yuan has a total of 41 co-authors. Here are the top co-authors by number of shared publications:


Latest Publications

  • Open Access

    SOLAR: Switchable Output Layer for Accuracy and Robustness in Once-for-All Training

    by Shaharyar Ahmed Khan Tareen, Lei Fan, Xiaojing Yuan, Qin Lin, Bin Hu

    Once-for-All (OFA) training enables a single super-net to generate multiple sub-nets tailored to diverse deployment scenarios, supporting flexible trade-offs among accuracy, robustness, and model-size without retraining. However, as the number of supported sub-nets increases, excessive parameter sharing in the backbone limits representational capacity, leading to degraded calibration and reduced overall performance. To address this, we propose SOLAR (Switchable Output Layer for Accuracy and Robustness in Once-for-All Training), a simple yet effective technique that assigns each sub-net a separate classification head. By decoupling the logit learning process across sub-nets, the Switchable Output Layer (SOL) reduces representational interference and improves optimization, without altering the shared backbone. We evaluate SOLAR on five datasets (SVHN, CIFAR-10, STL-10, CIFAR-100, and TinyImageNet) using four super-net backbones (ResNet-34, WideResNet-16-8, WideResNet-40-2, and MobileNetV2) for two OFA training frameworks (OATS and SNNs). Experiments show that SOLAR outperforms the baseline methods: compared to OATS, it improves accuracy of sub-nets up to 1.26 %, 4.71 %, 1.67 %, and 1.76 %, and robustness up to 9.01 %, 7.71 %, 2.72 %, and 1.26 % on SVHN, CIFAR-10, STL-10, and CIFAR-100, respectively. Compared to SNNs, it improves TinyImageNet accuracy by up to 2.93 %, 2.34 %, and 1.35 % using ResNet-34, WideResNet-16-8, and MobileNetV2 backbones (with 8 sub-nets), respectively.

  • Smart Guide Signing System Using RFID As an Intelligent Sensor

    by F. Qiao, L. Yu, R. Fatholahzadeh, X. Yuan

    Intersections account for more than two million crashes in the United States every year. Besides good geometric designs, an even more efficient and powerful way is to design a smart guide signing system to accurately and timely provide drivers with proper control and regulation information (e.g., STOP and YIELD signs, speed limit), which are however normally blocked by roadside trees and parking vehicles. The Radio Frequency Identification (RFID) technology, which is ideal in providing short range "infrastructure (i.e., the roadways) to vehicle" communications, is one of the good tools to implement such smart guide system. In this paper, a pilot test using RFID as an intelligent sensor is presented, including the selection of proper RFID emitters and receivers, the testing design, data collection from the tests and performance analysis, computer programming, simulation of in-vehicle display system to broadcast the received information, etc. Besides, the possible impacts of RFID based smart guide system to transportation safety, environment, and efficiency are also provided.


  • Implementation of the 7-point checklist for melanoma detection on smart handheld devices

    by Tarun Wadhawan, Ning Situ, Hu Rui, Keith Lancaster, Xiaojing Yuan, George Zouridakis

    In this paper we implement the 7-point checklist, a set of dermoscopic criteria widely used by clinicians for melanoma detection, on smart handheld devices, such as the Apple iPhone and iPad. The application developed is using sophisticated image processing and pattern recognition algorithms, yet it is light enough to run on a handheld device with limited memory and computational speed. When combined with a commercially available handheld dermoscope that provides proper lesion illumination, this application provides a truly self-contained handheld system for melanoma detection. Such a device can be used in a clinical setting for routine skin screening, or as an assistive diagnostic device in underserved areas and in developing countries with limited healthcare infrastructure.


  • A framework for regional association rule mining and scoping in spatial datasets

    by Wei Ding, Christoph F. Eick, Xiaojing Yuan, Jing Wang, Jean-Philippe Nicot


  • Introduce CDIO Model into Undergraduate Computer Engineering Curriculum—Networked Data Acquisition Systems

    by Xiaojing Yuan, Heidar Malki, Gangbing Song, Miguel Ramos

    With the rapid advancement in science, technology, engineering, and mathematics (STEM), it is getting more difficult to educate the future STEM professionals with emerging disciplinary knowledge. In recent years, the STEM education emphasizes more on students' learning outcomes, as can be seen from ABET, SACS, and other regional accreditation agencies. The CDIO model is a relatively new philosophy that adopted and emphasized the fundamentals in the context of product-system lifecycle from industry: Conceiving, Designing, Implementing, and Operating. In this paper, we detail the steps, procedures, instruments, and educational products we are developing to achieve the conceiving and operating of the CDIO model and present our thought on how the results may positively affect the outcome for the Designing and Implementing in the CDIO model based syllabus. The projects and experiments we created cover the fundamental knowledge of networked data acquisition systems, basic skills in designing and implementing such a system, and structured team-work.

  • Open Access

    Coi-wiz: An interactive computer wizard for analyzing cardiac optical signals

    by Xiaojing Yuan, Ilyas Uyanik, Ning Situ, Yutao Xi, Jie Cheng

    A number of revolutionary techniques have been developed for cardiac electrophysiology research to better study the various arrhythmia mechanisms that can enhance ablating strategies for cardiac arrhythmias. Once the three-dimensional high resolution cardiac optical imaging data is acquired, it is time consuming to manually go through them and try to identify the patterns associated with various arrhythmia symptoms. In this paper, we present an interactive computer wizard that helps cardiac electrophysiology researchers to visualize and analyze the high resolution cardiac optical imaging data. The wizard provides a file interface that accommodates different file formats. A series of analysis algorithms output waveforms, activation and action potential maps after spatial and temporal filtering, velocity field and heterogeneity measure. The interactive GUI allows the researcher to identify the region of interest in both the spatial and temporal domain, thus enabling them to study different heart chamber at their choice.

  • Open Access

    Cross-layer fault tolerant data aggregation for improved network delay in healthcare management applications

    by Xiaojing Yuan, Haoying Liu

    The escalation of American health care costs compels a new approach to manage chronic diseases. Wireless sensor networks (WSN) have been applied successfully in remote monitoring in military, aerospace, civil structure, and healthcare. However, existing wireless network framework cannot provide required quality of service (QoS) due to communication device failure, message loss caused by link error, collision, and hidden terminal for personalized disease management applications. In this paper, we present scalable network architecture and an operating mechanism that tolerates network structure changes caused by failure, with the application level data aggregation algorithm able to heal from the failure. We provide close form solutions that can achieve optimized network delay. Performance analysis was done to evaluate the significance of different nodes' failure in both homogeneous and heterogeneous sensor network and the effects of sensing and communication speed on failure impact in heterogeneous sensor networks.

  • Performance Analysis of Data Aggregation in Wireless Sensor Mesh Networks

    by Haoying Liu, Jian Shen, Xiaojing Yuan, Mequanint Moges

    In recent times the interest in wireless sensor mesh networks has grown considerably from personal, to local and metropolitan areas deployment. These networks consist of several mesh routers with minimal mobility and mesh clients that can be either mobile or stationary. The clients may also form a client mesh network among themselves and with routers. The various nodes over the network are interconnected via wireless links which might possibly employ multiple radio interfaces. One major attribute of such networks is the presence of redundant links which removes the single point failure that is present in the classical star or tree networks. Most researches in this field focus on the study of various routing protocols while we sought to introduce the application divisible load theory to find an optimum data aggregation strategy that optimize the networks performance with respect to response time and network delay. We define data aggregation as the process of data sensing and reporting back to the sink nodes, typically routers. The performance of wireless mesh network with 25 sensor nodes is examined by varying network bandwidth and sensing power of sensor nodes. Basic recursive equations for sensing and data reporting are developed for the case of homogeneous and heterogeneous mesh networks and the performance results of two representative data sensing and reporting strategies are presented.


  • A Fault Tolerant Efficient Scheduling Method for Improved Network Delay in Distributed Sensor Networks

    by Xiaojing Yuan, Haoying Liu, Jian Shen

    Wireless sensor networks are now applied in many fields which deal with data-centric information or critical mission that need minimized network delay. At same time, it needs the capability of dealing with faulty situations. In this work, we propose scalable network architecture and an operating mechanism that in nature is tolerant to network structure change caused by failure, and extend our application level scheduling algorithm with the capability of healing from the failure. The algorithm achieves optimized network delay and gets close form solution that is computational simple to implement. We also look into the significance of different nodes' failure in both homogeneous and heterogeneous sensor network. Simulation results imply the effects of sensing and communication speed on significance of failure in heterogeneous sensor networks.


  • DS-MAC: Differential service Medium Access Control design for wireless medical information systems

    by Xiaojing Yuan, Sumegha Bagga, Jian Shen, M. Balakrishnan, D. Benhaddou

    The integration of wireless networking technologies with medical information systems (telemedicine) have a significant impact on healthcare services provided to our society. Applications of telemedicine range from personalized medicine to affordable healthcare for underserved population. Though wireless technologies and medical informatics are individually progressing rapidly, wireless networking for healthcare systems is still at a very premature stage. In this paper we first present our open architecture for medical information systems that integrates both wired and wireless networked data acquisition systems. We then present the implementation at the physical layer and differential service MAC design that adapts channel provisioning based on the information criticality. Performance evaluation using analytical modeling and simulation shows that our DS-MAC provides differentiated services for emergency, warning, and normal traffic.