• The Journal of Korean Institute of Communications and Information Sciences
• /
• v.29 no.4C
• /
• pp.494-504
• /
• 2004

The researches about gaze detection have been much developed with many applications. Most previous researches only rely on image processing algorithm, so they take much processing time and have many constraints. In our work, we implement it with a computer vision system setting a IR-LED based single camera. To detect the gaze position, we locate facial features, which is effectively performed with IR-LED based camera and SVM(Support Vector Machine). When a user gazes at a position of monitor, we can compute the 3D positions of those features based on 3D rotation and translation estimation and affine transform. Finally, the gaze position by the facial movements is computed from the normal vector of the plane determined by those computed 3D positions of features. In addition, we use a trained neural network to detect the gaze position by eye's movement. As experimental results, we can obtain the facial and eye gaze position on a monitor and the gaze position accuracy between the computed positions and the real ones is about 4.2 cm of RMS error.

• The Journal of Korean Institute of Next Generation Computing
• /
• v.15 no.6
• /
• pp.7-14
• /
• 2019

The automatic monitoring and detection of crowd behavior in the surveillance videos has obtained significant attention in the field of computer vision due to its vast applications such as security, safety and protection of assets etc. Also, the field of crowd analysis is growing upwards in the research community. For this purpose, it is very necessary to detect and analyze the crowd behavior. In this paper, we proposed a deep learning-based method which detects abnormal activities in surveillance cameras installed in a smart city. A fine-tuned VGG-16 model is trained on publicly available benchmark crowd dataset and is tested on real-time streaming. The CCTV camera captures the video stream, when abnormal activity is detected, an alert is generated and is sent to the nearest police station to take immediate action before further loss. We experimentally have proven that the proposed method outperforms over the existing state-of-the-art techniques.

• Journal of the Korea Institute of Information and Communication Engineering
• /
• v.25 no.5
• /
• pp.637-644
• /
• 2021

In general, in the fields of computer vision, robotics, and augmented reality, the importance of 3D space and 3D object detection and recognition technology has emerged. In particular, since it is possible to acquire RGB images and depth images in real time through an image sensor using Microsoft Kinect method, many changes have been made to object detection, tracking and recognition studies. In this paper, we propose a method to improve the quality of 3D reconstructed images by processing images acquired through a depth-based (RGB-Depth) camera on a multi-view camera system. In this paper, a method of removing noise outside an object by applying a mask acquired from a color image and a method of applying a combined filtering operation to obtain the difference in depth information between pixels inside the object is proposed. Through each experiment result, it was confirmed that the proposed method can effectively remove noise and improve the quality of 3D reconstructed image.

• International conference on construction engineering and project management
• /
• 2020.12a
• /
• pp.313-322
• /
• 2020

The recognition of the risk hazards is a vital step to effectively prevent accidents on a construction site. The advanced development in computer vision systems and the availability of the large visual database related to construction site made it possible to take quick action in the event of human error and disaster situations that may occur during management supervision. Therefore, it is necessary to analyze the risk factors that need to be managed at the construction site and review appropriate and effective technical methods for each risk factor. This research focuses on analyzing Occupational Safety and Health Agency (OSHA) related to risk zone identification rules that can be adopted by the image recognition technology and classify their risk factors depending on the effective technical method. Therefore, this research developed a pattern-oriented classification of OSHA rules that can employ a large scale of safety hazard recognition. This research uses joint reasoning of risk zone Identification and numeric input by utilizing a stereo camera integrated with an image detection algorithm such as (YOLOv3) and Pyramid Stereo Matching Network (PSMNet). The research result identifies risk zones and raises alarm if a target object enters this zone. It also determines numerical information of a target, which recognizes the length, spacing, and angle of the target. Applying image detection joint logic algorithms might leverage the speed and accuracy of hazard detection due to merging more than one factor to prevent accidents in the job site.

• Journal of Internet Computing and Services
• /
• v.24 no.6
• /
• pp.81-89
• /
• 2023

In modern society, blind people face difficulties in navigating common environments such as sidewalks, elevators, and crosswalks. Research has been conducted to alleviate these inconveniences for the visually impaired through the use of visual and audio aids. However, such research often encounters limitations when it comes to practical implementation due to the high cost of wearable devices, high-performance CCTV systems, and voice sensors. In this paper, we propose an artificial intelligence fusion algorithm that utilizes low-cost video sensors integrated into smartphones to help blind people safely navigate their surroundings during walking. The proposed algorithm combines motion capture and object detection algorithms to detect moving people and various obstacles encountered during walking. We employed the MediaPipe library for motion capture to model and detect surrounding pedestrians during motion. Additionally, we used object detection algorithms to model and detect various obstacles that can occur during walking on sidewalks. Through experimentation, we validated the performance of the artificial intelligence fusion algorithm, achieving accuracy of 0.92, precision of 0.91, recall of 0.99, and an F1 score of 0.95. This research can assist blind people in navigating through obstacles such as bollards, shared scooters, and vehicles encountered during walking, thereby enhancing their mobility and safety.

• Journal of Intelligence and Information Systems
• /
• v.28 no.1
• /
• pp.89-106
• /
• 2022

Recently, the amount of video data collected from smartphones, CCTVs, black boxes, and high-definition cameras has increased rapidly. According to the increasing video data, the requirements for analysis and utilization are increasing. Due to the lack of skilled manpower to analyze videos in many industries, machine learning and artificial intelligence are actively used to assist manpower. In this situation, the demand for various computer vision technologies such as object detection and tracking, action detection, emotion detection, and Re-ID also increased rapidly. However, the object detection and tracking technology has many difficulties that degrade performance, such as re-appearance after the object's departure from the video recording location, and occlusion. Accordingly, action and emotion detection models based on object detection and tracking models also have difficulties in extracting data for each object. In addition, deep learning architectures consist of various models suffer from performance degradation due to bottlenects and lack of optimization. In this study, we propose an video analysis system consists of YOLOv5 based DeepSORT object tracking model, SlowFast based action recognition model, Torchreid based Re-ID model, and AWS Rekognition which is emotion recognition service. Proposed model uses single-linkage hierarchical clustering based Re-ID and some processing method which maximize hardware throughput. It has higher accuracy than the performance of the re-identification model using simple metrics, near real-time processing performance, and prevents tracking failure due to object departure and re-emergence, occlusion, etc. By continuously linking the action and facial emotion detection results of each object to the same object, it is possible to efficiently analyze videos. The re-identification model extracts a feature vector from the bounding box of object image detected by the object tracking model for each frame, and applies the single-linkage hierarchical clustering from the past frame using the extracted feature vectors to identify the same object that failed to track. Through the above process, it is possible to re-track the same object that has failed to tracking in the case of re-appearance or occlusion after leaving the video location. As a result, action and facial emotion detection results of the newly recognized object due to the tracking fails can be linked to those of the object that appeared in the past. On the other hand, as a way to improve processing performance, we introduce Bounding Box Queue by Object and Feature Queue method that can reduce RAM memory requirements while maximizing GPU memory throughput. Also we introduce the IoF(Intersection over Face) algorithm that allows facial emotion recognized through AWS Rekognition to be linked with object tracking information. The academic significance of this study is that the two-stage re-identification model can have real-time performance even in a high-cost environment that performs action and facial emotion detection according to processing techniques without reducing the accuracy by using simple metrics to achieve real-time performance. The practical implication of this study is that in various industrial fields that require action and facial emotion detection but have many difficulties due to the fails in object tracking can analyze videos effectively through proposed model. Proposed model which has high accuracy of retrace and processing performance can be used in various fields such as intelligent monitoring, observation services and behavioral or psychological analysis services where the integration of tracking information and extracted metadata creates greate industrial and business value. In the future, in order to measure the object tracking performance more precisely, there is a need to conduct an experiment using the MOT Challenge dataset, which is data used by many international conferences. We will investigate the problem that the IoF algorithm cannot solve to develop an additional complementary algorithm. In addition, we plan to conduct additional research to apply this model to various fields' dataset related to intelligent video analysis.

• The Journal of The Korea Institute of Intelligent Transport Systems
• /
• v.21 no.5
• /
• pp.274-286
• /
• 2022

Multi-object tracking has been studied for a long time under computer vision and plays a critical role in applications such as autonomous driving and driving assistance. Multi-object tracking techniques generally consist of a detector that detects objects and a tracker that tracks the detected objects. Various publicly available datasets allow us to train a detector model without much effort. However, there are relatively few publicly available datasets for training a tracker model, and configuring own tracker datasets takes a long time compared to configuring detector datasets. Hence, the detector is often developed separately with a tracker module. However, the separated tracker should be adjusted whenever the former detector model is changed. This study proposes a system that can train a model that performs detection and tracking simultaneously using only the detector training datasets. In particular, a Siam network with augmentation is used to compose the detector and tracker. Experiments are conducted on public datasets to verify that the proposed algorithm can formulate a real-time multi-object tracker comparable to the state-of-the-art tracker models.

• Journal of KIISE:Software and Applications
• /
• v.35 no.12
• /
• pp.780-788
• /
• 2008

Modeling hand poses and tracking its movement are one of the challenging problems in computer vision. There are two typical approaches for the reconstruction of hand poses in 3D, depending on the number of cameras from which images are captured. One is to capture images from multiple cameras or a stereo camera. The other is to capture images from a single camera. The former approach is relatively limited, because of the environmental constraints for setting up multiple cameras. In this paper we propose a method of reconstructing 3D hand poses from a 2D input image sequence captured from a single camera by means of Belief Propagation in a graphical model and recognizing a finger clicking motion using a hidden Markov model. We define a graphical model with hidden nodes representing joints of a hand, and observable nodes with the features extracted from a 2D input image sequence. To track hand poses in 3D, we use a Belief Propagation algorithm, which provides a robust and unified framework for inference in a graphical model. From the estimated 3D hand pose we extract the information for each finger's motion, which is then fed into a hidden Markov model. To recognize natural finger actions, we consider the movements of all the fingers to recognize a single finger's action. We applied the proposed method to a virtual keypad system and the result showed a high recognition rate of 94.66% with 300 test data.

• Journal of Digital Contents Society
• /
• v.15 no.4
• /
• pp.481-489
• /
• 2014

In this paper, the automatic standby power blocking socket outlet to reduce standby power by blocking power threshold is implemented. Where, the standby power means a flowing power when a disused power electronic is plugged into the socket outlet. The proposed socket outlet can cut off the standby power by establishing a proper block power threshold electronic device according to each electronic device because it can monitor the amount of power through the smart machines such as the real-time PC or mobile phone and directly control the blocking power threshold. The software is implemented by using Visual Studio software, code vision and SN8 C studio, and the hardware is embodied in ATmega128, SN8F27E93S, USB to UART, and relay etc. Through the simulation, we find that the standby power of the proposed method is similar to that of the conventional method in case of the cellular phone but the standby power of the proposed method is much less than that of the conventional method in case of the computer, air conditioning, and set-top box. Therefore, it is proved that the proposed socket outlet has a superior performance in terms of the standby power.

• Agricultural and Biosystems Engineering
• /
• v.4 no.1
• /
• pp.8-15
• /
• 2003

In Korea, quality evaluation of dried oak mushrooms are done first by classifying them into more than 10 different categories based on the state of opening of the cap, surface pattern, and colors. And mushrooms of each category are further classified into 3 or 4 groups based on its shape and size, resulting into total 30 to 40 different grades. Quality evaluation and sorting based on the external visual features are usually done manually. Since visual features of mushroom affecting quality grades are distributed over the entire surface of the mushroom, both front (cap) and back (stem and gill) surfaces should be inspected thoroughly. In fact, it is almost impossible for human to inspect every mushroom, especially when they are fed continuously via conveyor. In this paper, considering real time on-line system implementation, image processing algorithms utilizing artificial neural network have been developed for the quality grading of a mushroom. The neural network based image processing utilized the raw gray value image of fed mushrooms captured by the camera without any complex image processing such as feature enhancement and extraction to identify the feeding state and to grade the quality of a mushroom. Developed algorithms were implemented to the prototype on-line grading and sorting system. The prototype was developed to simplify the system requirement and the overall mechanism. The system was composed of automatic devices for mushroom feeding and handling, a set of computer vision system with lighting chamber, one chip microprocessor based controller, and pneumatic actuators. The proposed grading scheme was tested using the prototype. Network training for the feeding state recognition and grading was done using static images. 200 samples (20 grade levels and 10 per each grade) were used for training. 300 samples (20 grade levels and 15 per each grade) were used to validate the trained network. By changing orientation of each sample, 600 data sets were made for the test and the trained network showed around 91 % of the grading accuracy. Though image processing itself required approximately less than 0.3 second depending on a mushroom, because of the actuating device and control response, average 0.6 to 0.7 second was required for grading and sorting of a mushroom resulting into the processing capability of 5,000/hr to 6,000/hr.