Yoojin Chung

Predicting protein-protein interactions from one feature using SVM

One of the difficulties of using Artificial Neural Networks (ANNs) to estimate atmospheric temperature is the large number of potential input variables available. In this study, four different feature extraction methods were used to reduce the input vector to train four networks to estimate temperature at different atmospheric levels. The four techniques used were: genetic algorithms (GA), coefficient of determination (CoD), mutual information (MI) and simple neural analysis (SNA). The results demonstrate that of the four methods used for this data set, mutual information and simple neural analysis can generate networks that have a smaller input parameter set, while still maintaining a high degree of accuracy.

HORSIC: An efficient one-time signature scheme for wireless sensor networks

One-time signature schemes are promising candidates for broadcast authentication, which is an essential security primitive in wireless sensor networks. This paper proposes HORSIC, an efficient one-time signature scheme for broadcast authentication in wireless sensor networks. Different from previous one-time signature schemes, HORSIC can reduce the public key size and signature size simultaneously at the cost of increased overhead in key generation and signature verification.

Information Retrieval on an SCI-Based PC Cluster

This article presents an efficient parallel information retrieval (IR) system which provides fast information service for the Internet users on low-cost high-performance PC-NOW environment. The IR system is implemented on a PC cluster based on the scalable coherent interface (SCI), a powerful interconnecting mechanism for both shared memory models and message-passing models. In the IR system, the inverted-index file (IIF) is partitioned into pieces using a greedy declustering algorithm and distributed to the cluster nodes to be stored on each nodes hard disk. For each incoming users query with multiple terms, terms are sent to the corresponding nodes which contain the relevant pieces of the IIF to be evaluated in parallel. The IR system is developed using a distributed-shared memory (DSM) programming technique based on the SCI. According to the experiments, the IR system outperforms an MPI-based IR system using Fast Ethernet as an interconnect. Speed-up of up to 5.0 was obtained with an 8-node cluster in processing each query on a 500,000-document IIF.

PARALLEL MAXIMUM MATCHING ALGORITHMS IN INTERVAL GRAPHS

We develop efficient parallel maximum matching algorithms in interval graphs by exploiting special properties of interval graphs. For general interval graphs, our algorithm requires O(log2v+(n log n/v) time and O(nv2+n2) operations on the CREW PRAM, where n is the number of intervals and v≤n is a parameter. By choosing , we obtain an -time algorithm in O(n2) operations. For v=n/log n, we have an O(log2n)-time algorithm in O(n3/log2n) operations. The best previously known solution takes O(log2n) time using O(n3log2n) operations. For proper interval graphs, our algorithm runs in O(log n) time using O(n) operations if input intervals are sorted and using O(nlog n) operations otherwise on the EREW PRAM.

Tracking of mitochondrial transports using a particle filtering method with a spatial constraint

This paper describes a tracking method to trace the movements of fluorescently labeled mitochondria in time-lapse image sequences. It is based on particle filtering, which is a state-of-the-art tracking method, and is enhanced with a spatial constraint to improve robustness. Since mitochondria move only through axons, the spatial constraint is generated by axon segmentation on a single frame, which is the average of all the frames. The spatial constraint limits the search space of the state vector and, consequently, lowers the chance for the tracking to get lost. Using a background subtraction algorithm, the proposed method is also equipped with automatic detection of starting points, thus minimizing the requirement for user input. Implementation of the proposed method for tracking of fluorescently labeled mitochondria in time-lapse images showed substantially improved robustness and speed compared to a conventional method. With these improvements, this new particle tracking method is expected to increase the throughput of fluorescently labeled mitochondrial transport experiments, which are required for neuroscience research.

Virtual protocol stack interface for multiple wireless sensor network simulators

We present VPSI, which is a Virtual Protocol Stack Interface for wireless network simulators. VPSI provides two features; 1) a unified programming interface for developing network protocols, and 2) separation of simulator engine and the network protocols. By using VPSI, developers do not need to understand multiple different simulator engines. We implemented and adapted VPSI on existing well-known simulation packages, and we observed that VPSI works well with different wireless network simulators.

A Flexible Virtual Development Environment for Embedded Systems

On-time delivering of an embedded system solution to market is very crucial because the market is highly competitive and the demands of consumers rapidly change. Virtual development environment increases efficiency of the embedded system development because it enables developers to develop, execute, and verify an embedded system without real hardware. This paper deals with an implementation of a virtual development environment for ARM core-based embedded systems. The environment is developed based on ARMs ARMulator that is an instruction set simulation environment. The developed environment is extended to use SystemC hardware IPs by attaching a SystemC simulation engine to the modeled ASB bus. Therefore, the environment can use both ARMulator-based hardware models and SystemC-based hardware models. By adding hardware IP modules such as Memory controller, LCD controller, Interrupt controller, 1-ch DMA, UART, 2-ch Timer, Watchdog Timer, GPIO Ports and graphical user interface applications, the ARMulator environment is expanded to a virtual development environment for hand-held devices and general applications. In addition, a real-time operating system muC/OS-II is ported to the simulation environment so that the environment can be used to develop muC/OS-II-based application software. A three-task test program verifies the functionality of the hardware IP modules and muC/OS-II operations. Compared to other environments, its construction cost is very low and the environment can be easily modified according to a engineers needs.

A user-specified approach for image contrast enhancement

We propose a user-specified histogram equalization (UHE) approach, which provides user with a single variable to have the control over the enhancement procedure. By tuning this single parameter, user can achieve the desired enhanced look of the intended objects in an image.

An internet protocol stack for high-speed transmission in a non-OS environment

Todays embedded systems are required to have stronger wired/wireless communication capability. Due to strict limitations on how they utilize resources such as power, address space, and processing ability, network protocols for embedded systems are designed to make the best use of constrained resources. They are developed in a way that allows for support from a range of operating systems and thus utilizes platform-independent architectures. This paper presents an Internet protocol stack design for embedded systems that operate without an operating system. Our scheme schedules the transmission and the reception of the data packets and uses cross-layer optimization. We implemented the proposed scheme in a LTE network device. The overhead of our scheme is low and it subsequently meets the constraint of transmission speed in the next generation mobile communication environments.

Pairwise matching of spots in 2-DE images using hopfield network

Matching spots between two-dimensional electrophoresis (2-DE) images is a bottleneck in the automation of proteome analysis. Because the matching problem is an NP-hard problem, the solution is usually a heuristic approach or a neural network method. So a Hopfield neural network approach is applied to solve this problem. An energy function is designed to represent the similarity of spots together with its neighbor spots. Experiment showed that Hopfield neural network with appropriate energy function and dynamics could solve the matching problem of spots in 2-DE images.