JCJin-Chern Chiou

Jin-Chern Chiou

*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


Jin-Chern Chiou has a total of 99 co-authors. Here are the top co-authors by number of shared publications:


Latest Publications

  • Radiation-Hardening Techniques and In-Orbit Robustness Validation for Trapezoidal Dual-Axis Sun Sensors in LEO Environments

    by Mang Ou-Yang, Ching-I Tai, Guan-Yu Huang, Tse-Yu Cheng, Chang-Hsun Liu, Yu-Siou Liu, Jin-Chern Chiou, Chen-Yu Chan, Tung-Yun Hsieh, Chen-Tsung Lin, et al.

  • Cross-Domain Fall Detection and Activity Recognition Using Low-Resolution Infrared Sensors and Deep Learning

    by Jin-Chern Chiou, Ting-Yi Chen, Chin-Cheng Wu

    Falls in health and social care require privacy-preserving and cost-effective monitoring solutions to ensure safety. Low-resolution infrared (IR) array sensors provide a promising balance between invasiveness and information content; however, their performance can be strongly affected by ambient temperature variations and thermal interference. In this work, an environment-adaptive fall-detection system was developed to achieve cross-domain robustness and embedded real-time feasibility. A 24×32 MLX90640 IR array integrated with an ESP32 microcontroller captures thermal sequences at 8 fps. The processing pipeline includes temperature-based filtering, background removal based on a U-Net architecture for isolating human motion, Farneback optical flow extraction, and a lightweight CNN–BiLSTM–Attention network for spatiotemporal modeling. Experiments conducted across three thermally distinct indoor environments (11,148 samples, six activity classes) demonstrate that the U-Net-based background subtraction significantly enhances model robustness, achieving a six-class classification accuracy of 86.31% and reaching 98.27% in a simplified three-class (normal/falling/empty) configuration. Ablation and profiling analyses confirm the contribution of each processing stage while maintaining a compact model (≈0.83 M parameters, ≈3.17 MB) operating in real time at 8 fps on the ESP32–MLX90640 platform.

  • Gait-based Biometric System Using Pressure Sensing Mats and Machine Learning Algorithms

    by Jin-Chern Chiou, Ching Yen, Fei-En Wu

    The COVID-19 pandemic has emphasized the awareness of avoiding contact with shared or public devices, such as those used in traditional biometric systems. Common biometric systems, including fingerprint, palmprint, and iris recognition, require physical contact with the device, which increases the risk of contracting infectious diseases. As a result, non-contact biometric systems, such as gait recognition, may be increasingly important in the future. In this paper, we present an accurate gait recognition system that uses pressure sensing mats. Our proposed system employs high-density pressure sensing mats that significantly reduce computational complexity when compared to traditional gait recognition methods that use cameras. We acquired pressure distribution data from 30 subjects, including 19 males and 11 females, and developed an algorithmic framework that involves data preprocessing and classification to identify different subjects. We implemented five supervised machine learning models as classifiers, and our results indicate that the Convolutional Neural Networks (CNN) model performed the best, with a classification accuracy of 92.08%. Our study shows that the proposed gait recognition system is an effective non-contact biometric system that can distinguish different individuals with high accuracy. The use of pressure sensing mats reduces the risks associated with physical contact, making it a promising solution for biometric recognition in public spaces during the ongoing COVID-19 pandemic. In conclusion, our research contributes to the development of non-contact biometric systems and presents a viable solution for future applications.


  • Open Access

    Band-Selection of a Portal LED-Induced Autofluorescence Multispectral Imager to Improve Oral Cancer Detection

    by Yung-Jhe Yan, Nai-Lun Cheng, Chia-Ing Jan, Ming-Hsui Tsai, Jin-Chern Chiou, Mang Ou-Yang

    This aim of this study was to find effective spectral bands for the early detection of oral cancer. The spectral images in different bands were acquired using a self-made portable light-emitting diode (LED)-induced autofluorescence multispectral imager equipped with 365 and 405 nm excitation LEDs, emission filters with center wavelengths of 470, 505, 525, 532, 550, 595, 632, 635, and 695 nm, and a color image sensor. The spectral images of 218 healthy points in 62 healthy participants and 218 tumor points in 62 patients were collected in the ex vivo trials at China Medical University Hospital. These ex vivo trials were similar to in vivo because the spectral images of anatomical specimens were immediately acquired after the on-site tumor resection. The spectral images associated with red, blue, and green filters correlated with and without nine emission filters were quantized by four computing method, including summated intensity, the highest number of the intensity level, entropy, and fractional dimension. The combination of four computing methods, two excitation light sources with two intensities, and 30 spectral bands in three experiments formed 264 classifiers. The quantized data in each classifier was divided into two groups: one was the training group optimizing the threshold of the quantized data, and the other was validating group tested under this optimized threshold. The sensitivity, specificity, and accuracy of each classifier were derived from these tests. To identify the influential spectral bands based on the area under the region and the testing results, a single-layer network learning process was used. This was compared to conventional rules-based approaches to show its superior and faster performance. Consequently, four emission filters with the center wavelengths of 470, 505, 532, and 550 nm were selected by an AI-based method and verified using a rule-based approach. The sensitivities of six classifiers using these emission filters were more significant than 90%. The average sensitivity of these was about 96.15%, the average specificity was approximately 69.55%, and the average accuracy was about 82.85%.


  • Quantitative examination of early diabetes by light-emitting diodes light-induced pupillary light reflex

    by Y. J. Yan, Y. C. Tsai, M. L. Ko, N. C. Lee, J. C. Chiou, M. Ou-Yang

    This study investigated the abnormal pupillary light reflex in patients with early diabetes mellitus (DM) without retinopathy by using a custom-made noninvasive portable pupilometer. The pupilometer recorded and analyzed the pupillary light reflex. Two light intensities, 0.2 cd and 1.2 cd, and four wavelengths of stimulus light—white (400 nm–800 nm), red (640 ± 5 nm), green (534 ± 5 nm), and blue (470 ± 5 nm)—were used to stimulate the pupil for 10 ms. The pupillary response was recorded for 15 s. A total of 40 healthy people and 40 people with DM without retinopathy participated in the experiment at the National Taiwan University Hospital. The mean and standard deviation of DM duration were 4.5 years and 3.9 years. Of the 16 indices, the duration that pupil restores from its minimum size to half of its resting size (DRP), maximum pupil restoration velocity (MRV), and average restoration velocity (ARV) exhibited the most significant differences between the healthy people and those with DM. Compared with healthy participants, DRP was 16.33% higher, and MRV and ARV were 17.45% and 4.58% lower, respectively, in those with DM. This might be attributable to the sympathetic nervous system (SNS) controlling the dilator muscle during the dark-adapted period and relaxing the pupil; the SNS had few degenerated nerve endings in people with DM. The three aforementioned indices might be used to evaluate the severity of autonomic neuropathy in early DM.


  • Open Access

    A High-Efficiency Driver Circuit for a Gas-Sensor Microheater Based on a Switch-Mode DC-to-DC Converter

    by Tzu-Sen Yang, Jin-Chern Chiou

    Low power consumption is one of the critical factors for successful Internet of Things (IoT) applications. In such applications, gas sensors have become a main source of power consumption because energy conversion efficiency of the microheater is relative over a wide range of operating temperatures. To improve the energy-conversion efficiency of gas-sensor microheaters, this paper proposes integrated switch-mode DC-to-DC power converter technology which we compare with traditional driving methods such as pulse-width modulation and the linear mode. The results indicate that energy conversion efficiency with this proposed method remains over 90% from 150 °C to 400 °C when using a 3.0, 4.2 and 5.0 V power supply. Energy-conversion efficiency increases by 1–74% compared with results obtained using the traditional driving methods, and the sensing film still detects alcohol and toluene at 200 °C and 280 °C, respectively, with high energy conversion efficiency. These results show that the proposed method is useful and should be further developed to drive gas-sensor microheaters, and then integrated into the circuits of the complementary metal-oxide-semiconductor micro electro mechanical systems (CMOS-MEMS).


  • Open Access

    Sensitivity Enhancement of Acetone Gas Sensor using Polyethylene Glycol/Multi-Walled Carbon Nanotubes Composite Sensing Film with Thermal Treatment

    by Jin-Chern Chiou, Chin-Cheng Wu, Tse-Mei Lin

    There is a need to develop a chemiresistive gas sensor equipped with a thermostat over a wide area for the sensor, which can protect the sensor from the influence of ambient temperature due to the uniform temperature of the thermostat. In this paper, we demonstrated an acetone gas sensor based on a polyethylene glycol (PEG)/Multi-walled Carbon Nanotubes (MWCNTs) composite film, which was equipped with a thermostat. The sensor was operated at modest working temperatures for sensor sensitivity enhancement. The optimum design of the polyimide-based thermostat with widely uniform thermal distribution was investigated in detail. It was found that the temperature uniformity of the thermostat was achieved using double spiral geometry. The experimental results of the sensor response showed that the PEG/MWCNTs composite film with a moderate working temperature revealed a higher sensitivity than that without thermal treatment. Moreover, the sensing mechanisms of the PEG/MWCNTs composite gas sensor to acetone vapor were studied as well.


  • Open Access

    A Wearable and Wireless Gas-Sensing System Using Flexible Polymer/Multi-Walled Carbon Nanotube Composite Films

    by Jin-Chern Chiou, Chin-Cheng Wu

    In this study, an integrated flexible gas sensor was developed based on a polymer/multi-walled carbon nanotube composite film by using Bluetooth wireless communication/interface technology. Polymer/multi-walled carbon nanotube composite films were deposited over a polyimide flexible substrate for building a gas sensor array by using a drop-casting method. Sensor response was acquired through interdigitated electrodes and multi-channel sensor boards, which were linked to a Bluetooth wireless transceiver. Additionally, a double-spiral-shaped heater was built into the backside of the gas sensor array as a thermostat to protect it from the influence of ambient temperature. Multi-channel sensing responses were read on a display screen via a smartphone application (app). The advantages of this system include light weight, low cost, highly integrated sensors, wireless telecommunication, and real-time functioning. Thus, it is a promising candidate for deployment in a wearable gas-sensing system used to study air pollution.


  • Open Access

    A Wirelessly Powered Smart Contact Lens with Reconfigurable Wide Range and Tunable Sensitivity Sensor Readout Circuitry

    by Jin-Chern Chiou, Shun-Hsi Hsu, Yu-Chieh Huang, Guan-Ting Yeh, Wei-Ting Liou, Cheng-Kai Kuei

    This study presented a wireless smart contact lens system that was composed of a reconfigurable capacitive sensor interface circuitry and wirelessly powered radio-frequency identification (RFID) addressable system for sensor control and data communication. In order to improve compliance and reduce user discomfort, a capacitive sensor was embedded on a soft contact lens of 200 μm thickness using commercially available bio-compatible lens material and a standard manufacturing process. The results indicated that the reconfigurable sensor interface achieved sensitivity and baseline tuning up to 120 pF while consuming only 110 μW power. The range and sensitivity tuning of the readout circuitry ensured a reliable operation with respect to sensor fabrication variations and independent calibration of the sensor baseline for individuals. The on-chip voltage scaling allowed the further extension of the detection range and prevented the implementation of large on-chip elements. The on-lens system enabled the detection of capacitive variation caused by pressure changes in the range of 2.25 to 30 mmHg and hydration level variation from a distance of 1 cm using incident power from an RFID reader at 26.5 dBm.