• Journal of the Korean Society for Precision Engineering
• /
• v.24 no.4 s.193
• /
• pp.76-83
• /
• 2007

A new neural network classifier is proposed for the automatic real-time surface inspection of high-speed cold steel strips having 11 different types of defects. 46 geometrical and gray-level features are extracted for the defect classification. 3241 samples of Posco's Kwangyang steel factory are used for training and testing the neural network classifier. The developed classifier produces plausible 15% error rate which is much better than 20-30% error rate of human vision inspection adopted in most of domestic steel factories.

• KSII Transactions on Internet and Information Systems (TIIS)
• /
• v.11 no.2
• /
• pp.1118-1133
• /
• 2017

Computer vision-based human activity recognition (HAR) has become very famous these days due to its applications in various fields such as smart home healthcare for elderly people. A video-based activity recognition system basically has many goals such as to react based on people's behavior that allows the systems to proactively assist them with their tasks. A novel approach is proposed in this work for depth video based human activity recognition using joint-based motion features of depth body shapes and Deep Belief Network (DBN). From depth video, different body parts of human activities are segmented first by means of a trained random forest. The motion features representing the magnitude and direction of each joint in next frame are extracted. Finally, the features are applied for training a DBN to be used for recognition later. The proposed HAR approach showed superior performance over conventional approaches on private and public datasets, indicating a prominent approach for practical applications in smartly controlled environments.

• Proceedings of the KIEE Conference
• /
• 2007.10a
• /
• pp.117-118
• /
• 2007

A robot needs a human detection algorithm for interaction with a human. This paper proposes a method that finds people using a SVM (support vector machine) classifier and a stereo camera. Feature vectors of SVM are extracted by HoG (histogram of gradient) within images. After training extracted vectors from the clustered images, the SVM algorithm creates a classifier for human detection. Each candidate for a human in the image is generated by clustering of depth information from a stereo camera and the candidate is evaluated by the classifier. When compared with the existing method of creating candidates for a human, clustering reduces computational time. The experimental results demonstrate that the proposed approach can be executed in real time.

• Korean Medical Education Review
• /
• v.19 no.3
• /
• pp.129-133
• /
• 2017

Premed education in the college of medicine at the Catholic University of Korea aims to promote student creativity and excellence in accordance with the mission of the college: to have a sense of calling, leadership, and competency. The Catholic Medical College premed curriculum includes 75 credits which are composed of 65 credits for required courses and 10 credits for elective courses. It consists of courses in basic science, medical science, liberal arts and humanities (premedical OMNIBUS). It also involves community programs in 'Vision and Mission,' 'Leadership Training,' and 'Academic Conference.' In addition, students are allowed self-directed choice of their courses and learning for one quarter.

• Journal of the Korea Society of Computer and Information
• /
• v.21 no.12
• /
• pp.125-130
• /
• 2016

Automatic quality inspection by computer vision can be applied and give a solution to the pharmaceutical industry field. Pharmaceutical capsule can be easily affected by flaws like dents, cracks, holes, etc. In order to solve the quality inspection problem, it is required computationally efficient image processing technique like thresholding, boundary edge detection and segmentation and some automated systems are available but they are very expensive to use. In this paper, we have developed a dented capsule image processing technique using edge-based image segmentation, TLS(Total Least Squares) curve fitting technique and adopted low cost camera module for capsule image capturing. We have tested and evaluated the accuracy, training and testing time of the classification recognition algorithms like PCA(Principal Component Analysis), ICA(Independent Component Analysis) and SVM(Support Vector Machine) to show the performance. With the result, PCA, ICA has low accuracy, but SVM has good accuracy to use for classifying the dented capsule.

• Proceedings of the KAIS Fall Conference
• /
• 2004.11a
• /
• pp.234-239
• /
• 2004

This paper describes preliminary results we have obtained in developing a computer vision system based on active IR illumination for real time gaze tracking for interactive graphic display. Unlike most of the existing gaze tracking techniques, which often require assuming a static head to work well and require a cumbersome calibration process fur each person, our gaze tracker can perform robust and accurate gaze estimation without calibration and under rather significant head movement. This is made possible by a new gaze calibration procedure that identifies the mapping from pupil parameters to screen coordinates using the Generalized Regression Neural Networks (GRNN). With GRNN, the mapping does not have to be an analytical function and head movement is explicitly accounted for by the gaze mapping function. Furthermore, the mapping function can generalize to other individuals not used in the training. The effectiveness of our gaze tracker is demonstrated by preliminary experiments that involve gaze-contingent interactive graphic display.

• Journal of Korea Multimedia Society
• /
• v.24 no.10
• /
• pp.1319-1325
• /
• 2021

It is important to remove fog for accurate object recognition and detection during preprocessing because images taken in foggy adverse weather suffer from poor quality of images due to scattering and absorption of light, resulting in poor performance of various vision-based applications. This paper proposes an end-to-end deep learning-based single image de-fogging method using U-Net architecture. The loss function used in the algorithm is a loss function based on Mahalanobis distance with fog features, which solves the problem of domain shifts, and demonstrates superior performance by comparing qualitative and quantitative numerical evaluations with conventional methods. We also design it to generate fog through the VGG19 loss function and use it as the next training dataset.

• Proceedings of the Korean Society of Broadcast Engineers Conference
• /
• 2019.06a
• /
• pp.163-165
• /
• 2019

The task of image denoising is gaining popularity in the computer vision research field. Its main objective of restoring the sharp image from given noisy input is demanded in all image processing procedure. In this work, we treat the process of residual hair removal on faces images similar to the task of image denoising. In particular, our method removes the residual hair that presents on the frontal or profile face images and in-paints it with the relevant skin color. To achieve this objective, we employ a deep neural network that able to perform both tasks in one time. Furthermore, simple technic of residual hair color augmentation is introduced to increase the number of training data. This approach is beneficial for improving the robustness of the network. Finally, we show that the experimental results demonstrate the superiority of our network in both quantitative and qualitative performances.

• Journal of Information Processing Systems
• /
• v.17 no.3
• /
• pp.453-461
• /
• 2021

Many deep learning approaches have been studied for image classification in computer vision. However, there are not enough data to generate accurate models in medical fields, and many datasets are not annotated. This study presents a new method that can use both unlabeled and labeled data. The proposed method is applied to classify cervix images into normal versus cancerous, and we demonstrate the results. First, we use a patch self-supervised learning for training the global context of the image using an unlabeled image dataset. Second, we generate a classifier model by using the transferred knowledge from self-supervised learning. We also apply attention learning to capture the local features of the image. The combined method provides better performance than state-of-the-art approaches in accuracy and sensitivity.

• Journal of Information Processing Systems
• /
• v.18 no.5
• /
• pp.599-613
• /
• 2022

Non-proliferative diabetic retinopathy is a representative complication of diabetic patients and is known to be a major cause of impaired vision and blindness. There has been ongoing research on automatic detection of diabetic retinopathy, however, there is also a growing need for research on an automatic severity classification system. This study proposes an automatic detection system for pathological symptoms of diabetic retinopathy such as microaneurysms, retinal hemorrhage, and hard exudate by applying the Faster R-CNN technique. An automatic severity classification system was devised by training and testing a Random Forest classifier based on the data obtained through preprocessing of detected features. An experiment of classifying 228 test fundus images with the proposed classification system showed 97.8% accuracy.