Hyun-Soo Choi

Biometric Authentication Using Noisy Electrocardiograms Acquired by Mobile Sensors

Electrocardiogram (ECG) signals from mobile sensors are expected to increase the availability of authentication in the emerging wearable device industry. However, mobile sensors provide a relatively lower quality signal than the conventional medical devices. This paper proposes a practical authentication procedure for ECG signals that collected via one-chip-solution mobile sensors. We designed a cascading bandpass filter for noise cancellation and suggest eight fiducial features. For classification-based authentication, we use the radial basis function kernel-based support vector machine showing the best performance among nine classifiers through experimental comparisons. In spite of noisy ECG signals in mobile sensors, we achieved 4.61% of the equal error rate (EER) on a single heartbeat, and 1.87% of EER on 15 s testing time on 175 subjects, which is a reasonable result and supports the usability of low-cost ECGs for biometric authentication.

FingerNet: Deep learning-based robust finger joint detection from radiographs

Radiographic image assessment is the most common method used to measure physical maturity and diagnose growth disorders, hereditary diseases and rheumatoid arthritis, with hand radiography being one of the most frequently used techniques due to its simplicity and minimal exposure to radiation. Finger joints are considered as especially important factors in hand skeleton examination. Although several automation methods for finger joint detection have been proposed, low accuracy and reliability are hindering full-scale adoption into clinical fields. In this paper, we propose FingerNet, a novel approach for the detection of all finger joints from hand radiograph images based on convolutional neural networks, which requires little user intervention. The system achieved 98.02% average detection accuracy for 130 test data sets containing over 1,950 joints. Further analysis was performed to verify the system robustness against factors such as epiphysis and metaphysis in different age groups.

Gene pathways and cell cycle-related genes in cultured avian primordial germ cells

Primordial germ cells (PGC) from early embryos are applicable to various kinds of research, including the production of transgenic animals. Primordial germ cells eventually migrate and differentiate into germ cells in the gonads, where they settle and rapidly proliferate. However, the proliferation rate of PGC is low in early embryos, and there are many significant pathways that mediate PGC activity. Therefore, in vitro culture of PGC from early embryos with efficient growth factors has been necessary. Recently, we cultured chicken PGC from embryonic d 2.5 with basic fibroblast growth factor and characterized the PGC through analysis of cell morphology, survival, proliferation, and apoptosis. However, large-scale analyses of genes expressed in cultured PGC and the genes involved in associated pathways are limited. The objective of the present investigation was to identify the signaling and metabolic pathways of expressed genes by microarray comparison between PGC and their somatic counterpart, chicken embryonic fibroblasts (CEF). We identified 795 genes that were expressed more predominantly in PGC and 824 genes that were expressed more predominantly in CEF. Among the predominant genes in PGC, 201 were differentially identified in 106 pathways. Among the predominant genes in CEF, 242 were differentially identified in 99 pathways. To further validate the genes involved in at least one candidate pathway, those involved in the cell cycle (12 predominant genes in PGC and 8 predominant genes in CEF) were examined by real-time PCR. To the best of our knowledge, this study is the first to investigate signaling and metabolic pathways in cultured PGC.

XGBoost-Based Instantaneous Drowsiness Detection Framework Using Multitaper Spectral Information of Electroencephalography

The socioeconomic losses caused by extreme daytime drowsiness are enormous in these days. Hence, building a virtuous cycle system is necessary to improve work efficiency and safety by monitoring instantaneous drowsiness that can be used in any environment. In this paper, we propose a novel framework to detect extreme drowsiness using a short time segment (~ 2 s) of EEG which well represents immediate activity changes depending on a persons arousal, drowsiness, and sleep state. To develop the framework, we use multitaper power spectral density (MPSD) for feature extraction along with extreme gradient boosting (XGBoost) as a machine learning classifier. In addition, we suggest a novel drowsiness labeling method by combining the advantages of the psychomotor vigilance task and the electrooculography technique. By experimental evaluation, we show that the adopted MPSD and XGB techniques outperform other techniques used in previous studies. Finally, we identify that spectral components (theta, alpha, and gamma) and channels (Fp1, Fp2, T3, T4, O1, and O2) play an important role in our drowsiness detection framework, which could be extended to mobile devices.

Deep learning based low-cost high-accuracy diagnostic framework for dementia using comprehensive neuropsychological assessment profiles

BackgroundThe conventional scores of the neuropsychological batteries are not fully optimized for diagnosing dementia despite their variety and abundance of information. To achieve low-cost high-accuracy diagnose performance for dementia using a neuropsychological battery, a novel framework is proposed using the response profiles of 2666 cognitively normal elderly individuals and 435 dementia patients who have participated in the Korean Longitudinal Study on Cognitive Aging and Dementia (KLOSCAD).MethodsThe key idea of the proposed framework is to propose a cost-effective and precise two-stage classification procedure that employed Mini Mental Status Examination (MMSE) as a screening test and the KLOSCAD Neuropsychological Assessment Battery as a diagnostic test using deep learning. In addition, an evaluation procedure of redundant variables is introduced to prevent performance degradation. A missing data imputation method is also presented to increase the robustness by recovering information loss. The proposed deep neural networks (DNNs) architecture for the classification is validated through rigorous evaluation in comparison with various classifiers.ResultsThe k-nearest-neighbor imputation has been induced according to the proposed framework, and the proposed DNNs for two stage classification show the best accuracy compared to the other classifiers. Also, 49 redundant variables were removed, which improved diagnostic performance and suggested the potential of simplifying the assessment. Using this two-stage framework, we could get 8.06% higher diagnostic accuracy of dementia than MMSE alone and 64.13% less cost than KLOSCAD-N alone.ConclusionThe proposed framework could be applied to general dementia early detection programs to improve robustness, preciseness, and cost-effectiveness.