• Electronics and Telecommunications Trends
• /
• v.33 no.4
• /
• pp.33-42
• /
• 2018

Object classification and detection are fundamental technologies in computer vision and its applications. Recently, a deep-learning based approach has shown significant improvement in terms of object classification and detection. This report reviews the progress of deep-learning based object classification and detection in views of the ImageNet Large Scale Visual Recognition Challenge (ILSVRC), and analyzes recent trends of object classification and detection technology and its applications.

• The Journal of Korea Institute of Information, Electronics, and Communication Technology
• /
• v.1 no.3
• /
• pp.83-89
• /
• 2008

This paper introduces a visual object classification algorithm based on statistical information. Objects are characterized through the Histogram of Oriented Gradients (HOG) method and classification is performed using Multiclass AdaBoost. Salient features of an object's appearance are detected by HOG blocks Blocks of different sizes are tested to define the most suitable configuration. To select the most informative blocks for classification a multiclass AdaBoostSVM algorithm is applied. The proposed method has a high speed processing and classification rate. Results of the evaluation based on example of hand gesture recognition are presented.

• Proceedings of the KSRS Conference
• /
• 2003.11a
• /
• pp.1030-1032
• /
• 2003

This study a imed to analyze the effect of combined optical and radar image for the land cover classification in coastal region. The study area, Gyeonggi Bay area has one of the largest tidal ranges and has frequent land cover changes due to the several reclamations and rather intensive land uses. Ten land cover types were classified using several datasets of combining Landsat ETM+ and RADARSAT imagery. The synergic effects of the merged datasets were analyzed by both visual interpretation and an ordinary supervised classification. The merged optical and SAR datasets provided better discrimination among the land cover classes in the coastal area. The overall classification accuracy of merged datasets was improved to 86.5% as compared to 78% accuracy of using ETM+ only.

• Journal of the Korean Society for Aviation and Aeronautics
• /
• v.27 no.1
• /
• pp.26-33
• /
• 2019

It is necessary to establish a UAV pilot license and training system because the number of UAV-related accidents has rapidly risen. Most of accidents are caused by the human factors such as the lack of control skill and aviation knowledge. In this paper, we investigate licensing policy of small UAV pilots and examine the level of UAV licensing system and classification criteria based on comparative analysis of national cases such as USA, UK and China. Recently, the Ministry of Land, Infrastructure and Transport Affairs is planning to improve the safety regulation by taking into account the risk level of the licensing system, which has been classified according to the existing weight and commercial purpose. From the comparative analysis, we suggested a improvement policy for UAV licensing system in the view of pilot license segmentation, beyond Visual Line-of-sight flight and high risk UAV for non-commercial.

• PNF and Movement
• /
• v.22 no.2
• /
• pp.213-222
• /
• 2024

Purpose: The purpose of this study was to evaluate the effects of early body-weight-supported treadmill training on quadriceps strength, knee pain, and arthrogenic muscle inhibition (AMI) after knee surgery. Methods: Sixteen adults were selected. Seven patients in the experimental group performed body-weight-supported treadmill training, and nine patients in the control group performed general therapeutic exercise programs. Both groups received 20 minutes of neuromuscular electrical stimulation and 20 minutes of exercise therapy for two weeks. We measured quadriceps strength, visual analogue scale, and modified AMI classification grade before and after the intervention. Data were analyzed using the Mann-Whitney U test and Wilcoxon signed-rank test. Results: Within the experimental group, significant differences were observed in quadriceps strength, visual analogue scale, and modified AMI classification grade. Significant differences were observed between the before- and after-intervention groups in quadriceps strength and visual analogue scale. However, no significant differences were found in the modified AMI classification. Conclusion: The results of this study indicate that early body-weight-supported treadmill training may be an effective intervention for improving strength, reducing pain, and addressing arthrogenic muscle inhibition following knee surgery.

• Journal of the Korean Society of Clothing and Textiles
• /
• v.26 no.10
• /
• pp.1412-1423
• /
• 2002

The purpose of this study is to objectively explain the visual sensibility of clothing torture that satisfies the consumer's sensibility. The photo stimuli on clothing texture are divided into hard, soft transparent and brilliant. For the study of image 38 kinds of costume samples is used. The Study was measured by using Semantic Differential method. The subjects were 410 females in twenties. The data were analyzed by factor analysis, ANOVA, MDS and regression analysis. Data were analyzed by SPSS. The major findings of this research were as follows: 1. As a result of the factor analysis,5 factors of visual sensibility were consist of high qualities, touches, looks, lightness, and warmness or coolness.2. There were significant difference in visual sensibility based on classification of clothing texture.3. The clothing texture was classified as thin-full, flat-lumpy. 4. As a result of the regression analysis, preferences of consumers can be connected directly with buying behavior and satisfaction can be closely related with preferences and positive buying behavior.

• Proceedings of the Korean Institute of Interior Design Conference
• /
• 2001.05a
• /
• pp.121-127
• /
• 2001

An unified measure of user interface efficiency and aesthetics for cyber museum is proposed. First, general structure of cyber museum is discussed and hierarchical analyses are done for sample sites. Usability tests based on the hierarchical analyses yield statistics of user access frequency and persistency for each page, on which access probability is deduced. Second, visual occupancy, a measure of efficiency of user interface element based on access probability is defined. The hierarchical statistics of visual occupancy can be an index for characterization and classification of cyber museums. Examples are provided.

• The KIPS Transactions:PartB
• /
• v.18B no.6
• /
• pp.365-374
• /
• 2011

Recently, malicious video classification and filtering techniques are of practical interest as ones can easily access to malicious multimedia contents through the Internet, IPTV, online social network, and etc. Considerable research efforts have been made to developing malicious video classification and filtering systems. However, the malicious video classification and filtering is not still being from mature in terms of reliable classification/filtering performance. In particular, the most of conventional approaches have been limited to using only the spatial features (such as a ratio of skin regions and bag of visual words) for the purpose of malicious image classification. Hence, previous approaches have been restricted to achieving acceptable classification and filtering performance. In order to overcome the aforementioned limitation, we propose new malicious video classification framework that takes advantage of using both the spatial and temporal features that are readily extracted from a sequence of video frames. In particular, we develop the effective temporal features based on the motion periodicity feature and temporal correlation. In addition, to exploit the best data fusion approach aiming to combine the spatial and temporal features, the representative data fusion approaches are applied to the proposed framework. To demonstrate the effectiveness of our method, we collect 200 sexual intercourse videos and 200 non-sexual intercourse videos. Experimental results show that the proposed method increases 3.75% (from 92.25% to 96%) for classification of sexual intercourse video in terms of accuracy. Further, based on our experimental results, feature-level fusion approach (for fusing spatial and temporal features) is found to achieve the best classification accuracy.

• Korean Journal of Remote Sensing
• /
• v.21 no.5
• /
• pp.383-392
• /
• 2005

Vegetation classes, especially grass and tree classes, are often confused in classification when conventional spectral pattern recognition techniques are used to classify urban areas. This paper reports on a study to improve the classification results by using an automated process of considering height information in separating urban vegetation classes, specifically tree and grass, using three-band, high-spatial resolution, digital aerial imagery. Height information was derived photogrammetrically from stereo pair imagery using cross correlation image matching to estimate differential parallax for vegetation pixels. A threshold value of differential parallax was used to assess whether the original class was correct. The average increase in overall accuracy for three test stereo pairs was $7.8\%$, and detailed examination showed that pixels reclassified as grass improved the overall accuracy more than pixels reclassified as tree. Visual examination and statistical accuracy assessment of four test areas showed improvement in vegetation classification with the increase in accuracy ranging from $3.7\%\;to\;18.1\%$. Vegetation classification can, in fact, be improved by adding height information to the classification procedure.

• KSII Transactions on Internet and Information Systems (TIIS)
• /
• v.11 no.5
• /
• pp.2590-2606
• /
• 2017

Generally, research of classical image classification algorithms assume that training data and testing data are derived from the same domain with the same distribution. Unfortunately, in practical applications, this assumption is rarely met. Aiming at the problem, a domain adaption image classification approach based on multi-sparse representation is proposed in this paper. The existences of intermediate domains are hypothesized between the source and target domains. And each intermediate subspace is modeled through online dictionary learning with target data updating. On the one hand, the reconstruction error of the target data is guaranteed, on the other, the transition from the source domain to the target domain is as smooth as possible. An augmented feature representation produced by invariant sparse codes across the source, intermediate and target domain dictionaries is employed for across domain recognition. Experimental results verify the effectiveness of the proposed algorithm.