• The KIPS Transactions:PartB
• /
• v.9B no.3
• /
• pp.319-326
• /
• 2002

In this paper, we have designed and implemented cluttering classification systems- unsupervised classifiers-for the processing of satellite remote sensing images. Implemented systems adopt various design patterns which include a factory pattern and a strategy pattern to support various satellite images'formats and to design compatible systems. The clustering systems consist of sequential clustering, K-Means clustering, ISODATA clustering and Fuzzy C-Means clustering classifiers. The systems are tested by using a Landsat TM satellite image for the classification input. As results, these clustering systems are well designed to extract sample data for the classification of satellite images of which there is no previous knowledge. The systems can be provided with real-time base clustering tools, compatibilities and components' reusabilities as well.

• Journal of the Semiconductor & Display Technology
• /
• v.10 no.1
• /
• pp.113-118
• /
• 2011

This paper proposes computer visual inspection algorithms for PCB defects which are found in a manufacturing process. The proposed method can detect open circuit and short circuit on bare PCB without using any reference images. It performs adaptive threshold processing for the ROI (Region of Interest) of a target image, median filtering to remove noises, and then analyzes connected components of the binary image. In this paper, the connected components of circuit pattern are defined as 6 types. The proposed method classifies the connected components of the target image into 6 types, and determines an unclassified component as a defect of the circuit. The analysis of the original target image detects open circuits, while the analysis of the complement image finds short circuits. The machine vision inspection system is implemented using C language in an embedded Linux system for a high-speed real-time image processing. Experiment results show that the proposed algorithms are quite successful.

• Journal of Institute of Control, Robotics and Systems
• /
• v.4 no.6
• /
• pp.747-752
• /
• 1998

Ultrasonic sensors have been widely used to recognize the working environment for a mobile robot. However, their intrinsic problems, such as specular reflection, wide beam angle, and slow propagation velocity, require an excessive number of sensors to be integrated for achieving the sensing goal. This paper proposes a new measurement scheme which uses only two sets of ultrasonic sensors to determine the location and the type of a target surface. By measuring the time difference between the returned signals from the target surface, which are generated by two transmitters with 1 ㎳ difference, it classifies the type and determines the size of the target surface. Since the proposed sensor system uses only two sets of ultrasonic sensors to recognize and localize the target surface, it significantly simplifies the sensing system and reduces the signal processing time so that the working environment can be recognized in real time.

• Journal of Electrical Engineering and Technology
• /
• v.13 no.6
• /
• pp.2468-2478
• /
• 2018

We propose the Sparse Feature Convolutional Neural Network (SFCNN) to reduce the volume of convolutional neural networks (CNNs). Despite the superior classification performance of CNNs, their enormous network volume requires high computational cost and long processing time, making real-time applications such as online-training difficult. We propose an advanced network that reduces the volume of conventional CNNs by producing a region-based sparse feature map. To produce the sparse feature map, two complementary region-based value extraction methods, cluster max extraction and local value extraction, are proposed. Cluster max is selected as the main function based on experimental results. To evaluate SFCNN, we conduct an experiment with two conventional CNNs. The network trains 59 times faster and tests 81 times faster than the VGG network, with a 1.2% loss of accuracy in multi-class classification using the Caltech101 dataset. In vehicle classification using the GTI Vehicle Image Database, the network trains 88 times faster and tests 94 times faster than the conventional CNNs, with a 0.1% loss of accuracy.

• Journal of the Korea Academia-Industrial cooperation Society
• /
• v.7 no.6
• /
• pp.1160-1166
• /
• 2006

The information of vehicle is very important for maintaining traffic order under the present complex traffic environments. This paper proposes a new vehicle plate recognition mechanism that is essential to know the information of vehicle. The proposed method uses SVM which is excellent object classification compare to other methods. Two-class SVM is used to find the location of vehicle plate and multi-class SVM is used to recognize the characters in the plate. As a real-time processing system using multi-step image processing and recognition process this method recognizes several different vehicle plates. Through the experimental results of real environmental image and recognition using the proposed method, the performance is proven.

• Journal of IKEEE
• /
• v.27 no.3
• /
• pp.313-318
• /
• 2023

This paper proposes a neural network applying fine-tuning as a way to improve the performance of Video Classification based on Vision Transformer. Recently, the need for real-time video image analysis based on deep learning has emerged. Due to the characteristics of the existing CNN model used in Image Classification, it is difficult to analyze the association of consecutive frames. We want to find and solve the optimal model by comparing and analyzing the Vision Transformer and Non-local neural network models with the Attention mechanism. In addition, we propose an optimal fine-tuning neural network model by applying various methods of fine-tuning as a transfer learning method. The experiment trained the model with the UCF101 dataset and then verified the performance of the model by applying a transfer learning method to the UTA-RLDD dataset.

• The Journal of Korea Robotics Society
• /
• v.12 no.4
• /
• pp.425-431
• /
• 2017

As drones gain more popularity these days, drone detection becomes more important part of the drone systems for safety, privacy, crime prevention and etc. However, existing drone detection systems are expensive and heavy so that they are only suitable for industrial or military purpose. This paper proposes a novel approach for training Convolutional Neural Networks to detect drones from images that can be used in embedded systems. Unlike previous works that consider the class probability of the image areas where the class object exists, the proposed approach takes account of all areas in the image for robust classification and object detection. Moreover, a novel loss function is proposed for the CNN to learn more effectively from limited amount of training data. The experimental results with various drone images show that the proposed approach performs efficiently in real drone detection scenarios.

• Journal of Welding and Joining
• /
• v.17 no.2
• /
• pp.53-60
• /
• 1999

Conventionally, CCD camera and vision sensor using the projected pattern of light is generally used to inspect the weld bead defects. But with this method, a lot of time is needed for image preprocessing, stripe extraction and thinning, etc. In this study, laser vision sensor using the scanning beam of light is used to shorten the time required for image preprocessing. The software for deciding whether the weld bead is in proper shape or not in real time is developed. The criteria are based upon the classification of imperfections in metallic fusion welds(ISO 6520) and limits for imperfections(ISO 5817).

• Journal of IKEEE
• /
• v.24 no.4
• /
• pp.1141-1147
• /
• 2020

The hornet species are so similar in shape that they are difficult for non-experts to classify, and because the size of the objects is small and move fast, it is more difficult to detect and classify the species in real time. In this paper, we developed a system that classifies hornets species in real time based on a deep learning algorithm using a boundary box. In order to minimize the background area included in the bounding box when labeling the training image, we propose a method of selecting only the head and body of the hornet. It also experimentally compares existing boundary box-based object recognition algorithms to find the best algorithms that can detect wasps in real time and classify their species. As a result of the experiment, when the mish function was applied as the activation function of the convolution layer and the hornet images were tested using the YOLOv4 model with the Spatial Attention Module (SAM) applied before the object detection block, the average precision was 97.89% and the average recall was 98.69%.

• Journal of Korea Multimedia Society
• /
• v.25 no.1
• /
• pp.39-51
• /
• 2022

Image-based gender classification and age estimation of human are classic problems in computer vision. Most of researches in this field focus just only one task of either gender classification or age estimation and most of the reported methods for each task focus on accuracy performance and are not computationally light. Thus, running both tasks together simultaneously on low cost mobile or embedded systems with limited cpu processing speed and memory capacity are practically prohibited. In this paper, we propose a novel light-weight gender classification and age estimation method based on ensemble multitasking deep learning with light-weight processing neural network architecture, which processes both gender classification and age estimation simultaneously and in real-time even for embedded systems. Through experiments over various well-known datasets, it is shown that the proposed method performs comparably to the state-of-the-art gender classification and/or age estimation methods with respect to accuracy and runs fast enough (average 14fps) on a Jestson Nano embedded board.