• Journal of Information Processing Systems
• /
• v.18 no.3
• /
• pp.428-442
• /
• 2022

A reliable automatic passenger counting (APC) system is a key point in transportation related to the efficient scheduling and management of transport routes. In this study, we introduce a lightweight head detection network using deep learning applicable to an embedded system. Currently, object detection algorithms using deep learning have been found to be successful. However, these algorithms essentially need a graphics processing unit (GPU) to make them performable in real-time. So, we modify a Tiny-YOLOv3 network using certain techniques to speed up the proposed network and to make it more accurate in a non-GPU environment. Finally, we introduce an APC system, which is performable in real-time on embedded systems, using the proposed head detection algorithm. We implement and test the proposed APC system on a Samsung ARTIK 710 board. The experimental results on three public head datasets reflect the detection accuracy and efficiency of the proposed head detection network against Tiny-YOLOv3. Moreover, to test the proposed APC system, we measured the accuracy and recognition speed by repeating 50 instances of entering and 50 instances of exiting. These experimental results showed 99% accuracy and a 0.041-second recognition speed despite the fact that only the CPU was used.

• Journal of the Korea Society of Computer and Information
• /
• v.14 no.11
• /
• pp.179-186
• /
• 2009

In this paper, we propose a head detection method based on vertical and horizontal pixel histogram analysis in order to overcome drawbacks of the previous head detection approach using Haar-like feature-based face detection. In the proposed method, we create the vertical and horizontal foreground pixel histogram images from the background subtraction image, which represent the number of foreground pixels in the same vertical or horizontal position. Then we extract feature points of a head region by applying Harris corner detection method to the foreground pixel histogram images and by analyzing corner points. The proposal method shows robust head detection results even in the face image covering forelock by hairs or the back view image in which the previous approaches cannot detect the head regions.

• Smart Structures and Systems
• /
• v.18 no.5
• /
• pp.1063-1085
• /
• 2016

Taking advantage of the high sensitivity and long-distance detection capability of acoustic emission (AE) technique, this paper focuses on the crack detection in rail head, which is one of the most vulnerable parts of rail track. The AE source location and noise cancellation were studied on the basis of practical rail profile, material and operational noise. In order to simulate the actual AE events of rail head cracks, field tests were carried out to acquire the AE waves induced by pencil lead break (PLB) and operational noise of the railway system. Wavelet transform (WT) was first utilized to investigate the time-frequency characteristics and dispersion phenomena of AE waves. Here, the optimal mother wavelet was selected by minimizing the Shannon entropy of wavelet coefficients. Regarding the obvious dispersion of AE waves propagating along the rail head and the high operational noise, the wavelet transform-based modal analysis location (WTMAL) method was then proposed to locate the AE sources (i.e. simulated cracks) respectively for the PLB-induced AE signals with and without operational noise. For those AE signals inundated with operational noise, the Hilbert transform (HT)-based noise cancellation method was employed to improve the signal-to-noise ratio (SNR). Finally, the experimental results demonstrated that the proposed crack detection strategy could locate PLB-simulated AE sources effectively in the rail head even at high operational noise level, highlighting its potential for field application.

• Proceedings of the IEEK Conference
• /
• 2007.07a
• /
• pp.283-284
• /
• 2007

This paper proposes a method of real-time segmentation of moving region and detection of head position in a single omnidrectional camera Segmentation of moving region used background modeling method by a mixture of Gaussian(MOG) and shadow detection method. Circular constraint was proposed for detecting head position.

• International Journal of Fuzzy Logic and Intelligent Systems
• /
• v.16 no.2
• /
• pp.131-139
• /
• 2016

Recognition of human motions has become a main area of computer vision due to its potential human-computer interface (HCI) and surveillance. Among those existing recognition techniques for human motions, head detection and tracking is basis for all human motion recognitions. Various approaches have been tried to detect and trace the position of human head in two-dimensional (2D) images precisely. However, it is still a challenging problem because the human appearance is too changeable by pose, and images are affected by illumination change. To enhance the performance of head detection and tracking, the real-time three-dimensional (3D) data acquisition sensors such as time-of-flight and Kinect depth sensor are recently used. In this paper, we propose an effective feature extraction method, called adaptive local binary pattern (ALBP), for depth image based applications. Contrasting to well-known conventional local binary pattern (LBP), the proposed ALBP cannot only extract shape information without texture in depth images, but also is invariant distance change in range images. We apply the proposed ALBP for head detection and tracking in depth images to show its effectiveness and its usefulness.

• Journal of Korea Society of Digital Industry and Information Management
• /
• v.10 no.2
• /
• pp.75-82
• /
• 2014

MANET has various types of attacks. In particular, routing attacks using characteristics of movement of nodes and wireless communication is the most threatening because all nodes which configure network perform a function of router which forwards packets. Therefore, mechanisms that detect routing attacks and defense must be applied. In this paper, we proposed hierarchical structure attack detection techniques in order to improve the detection ability against routing attacks. Black hole detection is performed using PIT for monitoring about control packets within cluster and packet information management on the cluster head. Flooding attack prevention is performed using cooperation-based distributed detection technique by member nodes. For this, member node uses NTT for information management of neighbor nodes and threshold whether attack or not receives from cluster head. The performance of attack detection could be further improved by calculating at regular intervals threshold considering the total traffic within cluster in the cluster head.

• Proceedings of the Korean Society of Precision Engineering Conference
• /
• 2009.10a
• /
• pp.757-758
• /
• 2009

• Journal of Microbiology and Biotechnology
• /
• v.31 no.4
• /
• pp.610-620
• /
• 2021

There has been increasing interest in the head and neck squamous cell carcinoma (HNSCC) that is caused by high-risk human papillomavirus (HR-HPV) and has posed a significant challenge to Otolaryngologists. A rapid, sensitive, and reliable method is required for the detection of HR-HPV in clinical specimens to prevent and treat HPV-induced diseases. In this study, a multiple cross-linking spiral amplification (MCLSA) assay was developed for the visual detection of HPV-16. In the MCLSA assay, samples were incubated under optimized conditions at 62℃ for 45 min, and after mixing with the SYBR Green I (SGI) dye, the positive amplicons showed bright green fluorescence while the negative amplicons exhibited no obvious change. The specificity test revealed that the developed MCLSA technique had high specificity and could effectively distinguish all five HPV-16 strains from other pathogenic microorganisms. In terms of analytical sensitivity, the limit of detection (LoD) of MCLSA assay was approximately 5.4 × 101 copies/tube, which was 10-fold more sensitive than loop-mediated isothermal amplification (LAMP) and RT-PCR. The detection results of laryngeal cancer specimens collected from 46 patients with suspected HPV infection in the Liaoning region demonstrated that the positive detection rates of MCLSA and hybridized capture 2 kit were 32.61% (15/46). The true positive rate of the MCLSA assay was higher than that of RT-PCR (100% vs. 93.33%) and LAMP (100% vs. 86.67%). Therefore, the MCLSA assay developed in the present study could be a potentially useful tool for the point-of-care (PoC) diagnosis of HR-HPV, especially in resource-limited countries.

• Journal of Korea Multimedia Society
• /
• v.19 no.12
• /
• pp.1927-1935
• /
• 2016

This paper propose a framework for human body parts in RGB-D image. We conduct tasks of obtaining person area, finding candidate areas and local detection in order to detect hand, foot and head which have features of long accumulative geodesic distance. A person area is obtained with background subtraction and noise removal by using depth image which is robust to illumination change. Finding candidate areas performs construction of graph model which allows us to measure accumulative geodesic distance for the candidates. Instead of raw depth map, our approach constructs graph model with segmented regions by quadtree structure to improve searching time for the candidates. Local detection uses HOG based SVM for each parts, and head is detected for the first time. To minimize false detections for hand and foot parts, the candidates are classified with upper or lower body using the head position and properties of geodesic distance. Then, detect hand and foot with the local detectors. We evaluate our algorithm with datasets collected Kinect v2 sensor, and our approach shows good performance for head, hand and foot detection.

• The Journal of Korean Institute of Communications and Information Sciences
• /
• v.35 no.9C
• /
• pp.756-760
• /
• 2010

This paper is proposed to an novel facial behavior recognition system for driver's fatigue detection. Facial behavior is shown in various facial feature such as head expression, head pose, gaze, wrinkles. But it is very difficult to clearly discriminate a certain behavior by the obtained facial feature. Because, the behavior of a person is complicated and the face representing behavior is vague in providing enough information. The proposed system for facial behavior recognition first performs detection facial feature such as eye tracking, facial feature tracking, furrow detection, head orientation estimation, head motion detection and indicates the obtained feature by AU of FACS. On the basis of the obtained AU, it infers probability each state occur through Bayesian network.