• Journal of the Korean Institute of Intelligent Systems
• /
• v.13 no.2
• /
• pp.154-162
• /
• 2003

In this paper, Associative Memory with Cellular Neural Networks classifies and recognizes image patterns as an operator applied to image process. CNN processes nonlinear data in real-time like neural networks, and made by cell which communicates with each other directly through its neighbor cells as the Cellular Automata does. It is applied to the optimization problem, associative memory, pattern recognition, and computer vision. Image processing with CNN is appropriate to 2-D images, because each cell which corresponds to each pixel in the image is simultaneously processed in parallel. This paper shows the method for designing the structure of associative memory based on CNN and getting output image by choosing the most appropriate weight pattern among the whole learned weight pattern memories. Each template represents weight values between cells and updates them by learning. Hebbian rule is used for learning template weights and LMS algorithm is used for classification.

• The Journal of the Institute of Internet, Broadcasting and Communication
• /
• v.19 no.6
• /
• pp.227-233
• /
• 2019

Recently, along with the recent development of deep learning technique, neural networks are achieving success in computer vision filed. Convolutional neural network have shown outstanding performance in not only for a simple image classification task, but also for tasks with high difficulty such as object segmentation and detection. However many such deep learning models are based on supervised-learning, which requires more annotation labels than image-level label. Especially image semantic segmentation model requires pixel-level annotations for training, which is very. To solve these problems, this paper proposes a weakly-supervised semantic segmentation method which requires only image level label to train network. Existing weakly-supervised learning methods have limitations in detecting only specific area of object. In this paper, on the other hand, we use multi-classifier deep learning architecture so that our model recognizes more different parts of objects. The proposed method is evaluated using VOC 2012 validation dataset.

• Journal of the Korea Convergence Society
• /
• v.9 no.5
• /
• pp.1-6
• /
• 2018

In smart grid an accurate load forecasting is crucial in planning resources, which aids in improving its operation efficiency and reducing the dynamic uncertainties of energy systems. Research in this area has included the use of shallow neural networks and other machine learning techniques to solve this problem. Recent researches in the field of computer vision and speech recognition, have shown great promise for Deep Neural Networks (DNN). To improve the performance of daily electric peak load forecasting the paper presents a new deep neural network model which has the architecture of two multi-layer neural networks being serially connected. The proposed network model is progressively pre-learned layer by layer ahead of learning the whole network. For both one day and two day ahead peak load forecasting the proposed models are trained and tested using four years of hourly load data obtained from the Korea Power Exchange (KPX).

• 전자공학회논문지 IE
• /
• v.49 no.2
• /
• pp.30-39
• /
• 2012

An effective drowsy driver detection system is needed, because the probability of accident is high for drowsy driving and its severity is high at the time of accident. However, the drowsy driver detection system that uses bio-signals or vision is difficult to be utilized due to high cost. Thus, this paper proposes a drowsy driver detection algorithm by using steering angle sensor, which is attached to the most of vehicles at no additional cost, and vehicle information such as brake switch, throttle position signal, and vehicle speed. The proposed algorithm is based on jerk criterion, which is one of drowsy driver's steering patterns. In this paper, threshold value of each variable is presented and the proposed algorithm is evaluated by using acquired vehicle data from hardware in the loop simulation (HILS) through CAN communication and MATLAB program.

• Journal of the Institute of Electronics and Information Engineers
• /
• v.51 no.10
• /
• pp.190-196
• /
• 2014

This paper proposes a mobile-optimized pedestrian detection method using Cascade of HOG(Histograms of Oriented Gradients) for ADAS(Advanced Driver Assistance System) on smartphones. In order to use the limited resource of mobile platforms efficiently, the method is implemented by the OpenCL(Open Computing Language) library, and its processing time is reduced in the following two aspects. Firstly, the method sets a program build option specifically and adjusts work group sizes as variety of kernels in the host code. Secondly, it utilizes local memory and a LUT(Look-Up Table) in the kernel code to accelerate the program. For performance evaluation, the developed algorithm is compared with the mobile CPU-based OpenCV(Open Computer Vision) for Android function. The experimental results show that the processing speed is 25% faster than the OpenCV hogcascade.

• Journal of the Institute of Electronics Engineers of Korea CI
• /
• v.45 no.4
• /
• pp.34-40
• /
• 2008

This paper presents a vision based walking robot control system using hand shape recognition. To recognize hand shapes, the accurate hand boundary needs to be tracked in image obtained from moving camera. For this, we use an active contour model-based tracking approach with mean shift which reduces dependency of the active contour model to location of initial curve. The proposed system is composed of four modules: a hand detector, a hand tracker, a hand shape recognizer and a robot controller. The hand detector detects a skin color region, which has a specific shape, as hand in an image. Then, the hand tracking is performed using an active contour model with mean shift. Thereafter the hand shape recognition is performed using Hue moments. To assess the validity of the proposed system we tested the proposed system to a walking robot, RCB-1. The experimental results show the effectiveness of the proposed system.

• Journal of IKEEE
• /
• v.14 no.3
• /
• pp.224-231
• /
• 2010

Automatic detection of objects in images has been one of core challenges in the areas such as computer vision and pattern analysis. Especially, with the recent deployment of personal mobile devices such as smart phone, such technology is required to be transported to them. Usually, these smart phone users are equipped with devices such as camera, GPS, and gyroscope and provide various services through user-friendly interface. However, the smart phones fail to give excellent performance due to limited system resources. In this paper, we propose a new scheme to improve object recognition performance based on pre-computation and simple local features. In the pre-processing, we first find several representative parts from similar type objects and classify them. In addition, we extract features from each classified part and train them using regression functions. For a given query image, we first find candidate representative parts and compare them with trained information to recognize objects. Through experiments, we have shown that our proposed scheme can achieve resonable performance.

• The Journal of The Korea Institute of Intelligent Transport Systems
• /
• v.2 no.1 s.2
• /
• pp.75-84
• /
• 2003

Current techniques for road-traffic monitoring rely on sensors which have limited capabilities, are costly and disruptive to install. The use of video cameras coupled with computer vision techniques offers an attractive alternative to current sensors. Video based traffic monitoring systems are now being considered key points of advanced traffic management systems. In this paper, we propose the new method which extract the traffic information using video camera. The proposed method uses an adaptive updating scheme for background in order to reduce the false alarm rate due to various noises in images. also, the proposed extraction method of traffic information calculates the traffic volume ratio of vehicles passing through predefined detection area, which is defined by the length of profile occupied by cars over that of overall detection area. Then the ratio is used to define 8 different states of traffic and to interpret the state of vehicle flows. The proposed method is verified by an experiment using CCTV traffic data from urban area.

• Journal of the Institute of Electronics Engineers of Korea TC
• /
• v.43 no.9 s.351
• /
• pp.135-141
• /
• 2006

In order to prevent traffic accident, it is very important that the driver regulates the location of rear view mirror using the automatic seat regulation system which guarantees the maximum vision of the possibility for accuracy. In order to solve this problem the paper deals with the automatic seat control system which guarantees comfortable and safe seating and good visual field. Also a automatic car seat control algorithm has been developed to regulate the back mirror. Particularly, the automatic seat control algorithm function for the air bag operation in case of an accident has been added depending on passengers weight. Moreover when the driver passes a dangerous area an algorithm has been developed which gives the driver a naming sign and has been simulated in a ubiquitous environment. The simulation result proved that the Intelligence analysis for traffic accidents can reduce franc accidents more than 25% than the currently existing methods.

• Journal of the Institute of Electronics Engineers of Korea CI
• /
• v.41 no.4
• /
• pp.85-94
• /
• 2004

Object tracking in a real time image is one of interesting subjects in computer vision and many Practical application fields during the past couple of years. But sometimes existing systems cannot find all objects by recognizing background noise as object. This paper proposes a method of object detection and tracking using adaptive background image in real time. To detect object which is not influenced by illumination and to remove noise in background image, this system generates adaptive background image by real time background image updating. This system detects object using the difference between background image and input image from camera. After setting up Minimum Bounding Rectangle(MBR) using the internal point of detected object, the system tracks object through this MBR In addition, this paper evaluates the test result about performance of proposed method as compared with existing tracking algorithm.