• Journal of Korea Multimedia Society
• /
• v.5 no.5
• /
• pp.505-514
• /
• 2002

In this Paper, we propose a method of content-based image classification using neural network. The images for classification ate object images that can be divided into foreground and background. To deal with the object images efficiently, object region is extracted with a region segmentation technique in the preprocessing step. Features for the classification are texture and shape features extracted from wavelet transformed image. The neural network classifier is constructed with the extracted features and the back-propagation learning algorithm. Among the various texture features, the diagonal moment was more effective. A test with 300 training data and 300 test data composed of 10 images from each of 30 classes shows correct classification rates of 72.3% and 67%, respectively.

• Journal of Electrical Engineering and Technology
• /
• v.9 no.3
• /
• pp.1060-1071
• /
• 2014

Signals of the Electroencephalogram (EEG) can reflect the electrical background activity of the brain generated by the cerebral cortex nerve cells. This has been the mostly utilized signal, which helps in effective analysis of brain functions by supervised learning methods. In this paper, an approach for improving the accuracy of EEG signal classification is presented to detect epileptic seizures. Moreover, Independent Component Analysis (ICA) is incorporated as a preprocessing step and Short Time Fourier Transform (STFT) is used for denoising the signal adequately. Feature extraction of EEG signals is accomplished on the basis of three parameters namely, Standard Deviation, Correlation Dimension and Lyapunov Exponents. The Artificial Neural Network (ANN) is trained by incorporating Levenberg-Marquardt(LM) training algorithm into the backpropagation algorithm that results in high classification accuracy. Experimental results reveal that the methodology will improve the clinical service of the EEG recording and also provide better decision making in epileptic seizure detection than the existing techniques. The proposed EEG signal classification using feed forward Backpropagation Neural Network performs better than to the EEG signal classification using Adaptive Neuro Fuzzy Inference System (ANFIS) classifier in terms of accuracy, sensitivity, and specificity.

• KSCE Journal of Civil and Environmental Engineering Research
• /
• v.26 no.1D
• /
• pp.149-156
• /
• 2006

Existing method of landuse classification using aerial photographs or field survey requires relatively higher amount of time and cost due to necessary manual work. Especially in urban area where the pattern of landuse is densely aggregated, a landuse classification using satellite image is more complex. In this background, this study proposes a landuse classification method to utilize 1:1,000 digital topographic data and IKONOS satellite image. To prove the possibility of this method, the method was applied to Seoul metropolitan area. The results shows the total accuracy of approximately 95% and 14 landuse classes extracted. Based on the results from the pilot study, this method is applicable to landuse classification in urban area.

• Safety and Health at Work
• /
• v.11 no.2
• /
• pp.152-158
• /
• 2020

Background: The Globally Harmonized System of Classification and Labeling of Chemicals (GHS) was developed to enhance chemical classification and hazard communication systems worldwide. However, some of the elements such as building blocks and data sources have the potential to cause "disharmony" to the GHS, particularly in its classification results. It is known that some countries have developed their own lists of classified chemicals in accordance with the GHS to "standardize" the classification results within their respective countries. However, the lists of classified chemicals may not be consistent among these countries. Method: In this study, the lists of classified chemicals developed by the European Union, Japan, Malaysia, and New Zealand were selected for comparison of classification results for carcinogenicity, germ cell mutagenicity, and reproductive toxicity. Results: The findings show that only 54%, 66%, and 37% of the classification results for each Carcinogen, Mutagen and Reproductive toxicants hazard classes, respectively are the same among the selected countries. This indicates a "moderate" level of consistency among the classified chemicals lists. Conclusion: By using classification results for the carcinogenicity, germ cell mutagenicity, and reproductive toxicity hazard classes, this study demonstrates the "disharmony" in the classification results among the selected countries. We believe that the findings of this study deserve the attention of the relevant international bodies.

• Journal of the Korean Society for information Management
• /
• v.29 no.2
• /
• pp.205-224
• /
• 2012

One of the limitations of BOW method is that each term is recognized only by its form, failing to represent the term's meaning or thematic background. To overcome the limitation, different profiles for each term were defined by thematic categories depending on contextual characteristics. In this study, a specific term was used as a classification feature based on its meaning or thematic background through the process of comparing the context in those profiles with the occurrences in an actual document. The experiment was conducted in three phases; term weighting, ensemble classifier implementation, and feature selection. The classification performance was enhanced in all the phases with the ensemble classifier showing the highest performance score. Also, the outcome showed that the proposed method was effective in reducing the performance bias caused by the total number of learning documents.

• Journal of the Korea Institute of Information and Communication Engineering
• /
• v.18 no.11
• /
• pp.2721-2726
• /
• 2014

In a MOG algorithm using the GMM to subtract background, the model parameter computation and the object classification to be performed at every pixel require a huge computation and are the chief obstacles to its uses. This paper proposes a fast MOG algorithm that partly adopts the simple model parameter computation and the object classification skip on the basis of the object prediction. The former is applied to the pixels that gives little effect on the model parameter and the latter is applied to the pixels whose object prediction is firmly trusted. In comparative experiment between the conventional and proposed algorithms using videos, the proposed algorithm carries out the simple model parameter computation and the object classification skip over 77.75% and 92.97%, respectively, nevertheless it retains more than 99.98% and 99.36% in terms of image and moving object-unit average classification accuracies, respectively.

• The KIPS Transactions:PartB
• /
• v.15B no.1
• /
• pp.1-8
• /
• 2008

Generally, moving objects in surveillance video are extracted by background subtraction or frame difference method. However, moving cast shadows on object distort extracted figures which cause serious detection problems. Especially, analyzing vehicle information in video frames from a fixed surveillance camera on road, we obtain inaccurate results by shadow which vehicle causes. So, Shadow Elimination is essential to extract right objects from frames in surveillance video. And we use shadow removal algorithm for vehicle classification. In our paper, as we suppress moving cast shadow in object, we efficiently discriminate vehicle types. After we fit new object of shadow-removed object as three dimension object, we use extracted attributes for supervised learning to classify vehicle types. In experiment, we use 3 learning methods {IBL, C4.5, NN(Neural Network)} so that we evaluate the result of vehicle classification by shadow elimination.

• Journal of the Korean earth science society
• /
• v.37 no.1
• /
• pp.40-51
• /
• 2016

The problem has been pointed out that the domestic design response spectrum does not reflect site amplification, particularly in the high frequency bands, including the fact that site specific response spectrum from the observed ground motions appears relatively higher than design response spectrum. Among various methods, this study applied H/V spectral ratio of ground motion for estimating site amplification. This method, originated from S waves and Rayleigh waves, recently has been extended to Coda waves and background noise for estimating site amplification. For limited time of periods, 4 electric substation sites had operated seismic stations at two separate locations (bedrock and borehole) within each substation site. H/V spectral ratio of S wave, Coda wave, and background noise, was applied to 36 accelerations of 3 macro earthquakes (Odaesan, Jeju and Gongju earthquakes), larger than magnitude 3.4. observed simultaneously at each bedrock location within 4 electric substation sites. Site amplifications at the bedrock location of 4 sites were compared among S wave, Coda wave energy, and background noise, and then compared to the previous results from the borehole location data. The site classification was also tried using resonancy frequency information at each site and location. The results suggested that all the electric substation sites showed similar site amplification patterns among S wave, Coda wave, and background noise. Each station showed its own characteristics of site amplification property in low, high and specific resonance frequency ranges. Comparison of this study to other results using different method can give us much more information about dynamic amplification of domestic sites characteristics and site classification.

• Journal of the Korean Society of Surveying, Geodesy, Photogrammetry and Cartography
• /
• v.40 no.1
• /
• pp.23-30
• /
• 2022

Recently, there are various studies underway on the deep learning-used image matting methods. Even in the field of photogrammetry, a process of extracting information about relics from images photographed is essential to produce a high-quality realistic model. Such a process requires a great deal of time and manpower, so chroma-key has been used for extraction so far. This method is low in accuracy of sub-classification, however, it is difficult to apply the existing method to high-quality realistic models. Thus, this study attempted to remove background information from high-resolution relic images by using prior background information and trained learning data and evaluate both qualitative and quantitative results of the relic images extracted. As a result, this proposed method with FBA(manual trimap) showed quantitatively better results, and even in the qualitative evaluation, it was high in accuracy of classification around relics. Accordingly, this study confirmed the applicability of the proposed method in the indoor relic photography since it showed high accuracy and fast processing speed by acquiring prior background information when classifying high-resolution relic images.

• Journal of IKEEE
• /
• v.23 no.4
• /
• pp.1415-1422
• /
• 2019

Background subtraction is the major task of computer vision and image processing to detect changes in video. The best performing background subtraction is computationally expensive that cannot be used in real time in a typical computing environment. The proposed algorithm improves the background subtraction algorithm of the widely used MOG with the image resizing algorithm. The proposed image resizing algorithm is designed to drastically reduce the amount of computation and to utilize local information, which is robust against noise such as camera movement. Experimental results of the proposed algorithm have a classification capability that is close to the state of the art background subtraction method and the processing speed is more than 10 times faster.