• Journal of the Korean Institute of Intelligent Systems
• /
• v.20 no.3
• /
• pp.422-427
• /
• 2010

Korean Game industry, especially MMORPG(Massively Multiplayer Online Game) has been rapidly expanding in these days. But As game industry is growing, lots of online game security incidents have also been increasing and getting prevailing. One of the most critical security incidents is 'Game Bots', which are programs to play MMORPG instead of human players. If player let the game bots play for them, they can get a lot of benefic game elements (experience points, items, etc.) without any effort, and it is considered unfair to other players. Plenty of game companies try to prevent bots, but it does not work well. In this paper, we propose a behavior pattern model for detecting bots. We analyzed behaviors of human players as well as bots and identified six game features to build the model to differentiate game bots from human players. Based on these features, we made a Naive Bayesian classifier to reasoning the game bot or not. To evaluated our method, we used 10 game bot data and 6 human Player data. As a result, we classify Game bot and human player with 88% accuracy.

• Journal of the Korean Association of Geographic Information Studies
• /
• v.5 no.1
• /
• pp.27-37
• /
• 2002

The classification accuracy of land cover has been considered as one of the major issues to estimate pollution loads generated from diffuse landuse patterns in a watershed. This research aimed to assess the effects of the acquisition methods and sampling size of training reference data on the classification accuracy of land cover using an imagery acquired by optical sensor(OPS) on JERS-1. Two kinds of data acquisition methods were considered to prepare training data. The first was to assign a certain land cover type to a specific pixel based on the researchers subjective discriminating capacity about current land use and the second was attributed to an aerial photograph incorporated with digital maps with GIS. Three different sizes of samples, 0.3%, 0.5%, and 1.0% of all pixels, were applied to examine the consistency of the classified land cover with the training data of corresponding pixels. Maximum likelihood scheme was applied to classify the land use patterns of JERS-1 imagery. Classification run applying an aerial photograph achieved 18 % higher consistency with the training data than the run applying the researchers subjective discriminating capacity. Regarding the sample size, it was proposed that the size of training area should be selected at least over 1% of all of the pixels in the study area in order to obtain the accuracy with 95% for JERS-1 satellite imagery on a typical small-to-medium-size urbanized area.

• Journal of the Institute of Electronics and Information Engineers
• /
• v.53 no.3
• /
• pp.114-123
• /
• 2016

The effective feature extraction method for unmanned aerial vehicle (UAV) detection is proposed and verified in this paper. The UAV engine sound is harmonic complex tone whose frequency ratio is integer and its variation is continuous in time. Using these characteristic, we propose the feature vector composed of a mean and standard deviation of difference value between fundamental frequency with 1st overtone as well as mean variation of their frequency. It was revealed by simulation that the suggested feature vector has excellent discrimination in target signal identification from various interfering signals including frequency variation with time. By comparing Fisher scores, three features based on frequency show outstanding discrimination of measured UAV signals with low signal to noise ratio (SNR). Detection performance with simulated interference signal is compared by MFCC by using ELM classifier and the suggested feature vector shows 37.6% of performance improvement As the SNR increases with time, the proposed feature can detect the target signal ahead of MFCC that needs 4.5 dB higher signal power to detect the target.

• Korean Journal of Remote Sensing
• /
• v.25 no.2
• /
• pp.145-154
• /
• 2009

High spatial resolution satellite image classification has a limitation when only using the spectral information due to the complex spatial arrangement of features and spectral heterogeneity within each class. Therefore, the extraction of the spatial information is one of the most important steps in high resolution satellite image classification. This study proposes a new spatial feature extraction method, named SSI(Shape-Size Index). SSI uses a simple region-growing based image segmentation and allocates spatial property value in each segment. The extracted feature is integrated with spectral bands to improve overall classification accuracy. The classification is achieved by applying a SVM(Support Vector Machines) classifier. In order to evaluate the proposed feature extraction method, KOMPSAT-2 and QuickBird-2 data are used for experiments. It is demonstrated that proposed SSI algorithm leads to a notable increase in classification accuracy.

• Korean Journal of Remote Sensing
• /
• v.12 no.2
• /
• pp.169-198
• /
• 1996

The objectives of this investigation are : 1. To analyze spectral signature and the associated vegetation index for geometric illumination conditions inf1uenced by low solar elevation and high slope orientations in mountainous forest. 2. To assess the accuracy of the spectral angle mapper classification for the a winter land cover in comparison with the maximum likelihood classification. 3. To produce the image of water quality and water properties that could be used to estimate the water pollution sources and the tide-included by turbid water in estuarine and coastal areas. These objectives are to characterize environmental and ecological monitoring applications of the Nak-Dong River Basin by using Fuyo-1 OPS VNIR data acquired on December 26, 1992. The results of this paper are as follows : 1. The spectral digital numbers and vegetation indexes (NDVI and TVI) of mountainous forest are higher on the slope facing the sun than on the slope hidden the sun under low sun elevation condition. 2. The spectral angle mapper algorithm produces a more accurate land cover classification of areas with steep slope, various aspects and low solar elevation than the maximum likelihood classifier. 3. The maximum likelihood classification images can be used for identifying the location and movement of both freshwater and salt water, regardless of geometric illumination conditions. 4. The color-coded density sliced image of selected water bodies by using the near-infrared band 3 can provide distribution of the water quality of the Lower Nak-Dong River. 5. The color-coded normalized difference vegetation index image of the selected mountain forest is suitable to classify winter vegetation cover types, i.e., forest canopy densities for slope orientations.

• Korean Journal of Remote Sensing
• /
• v.21 no.3
• /
• pp.189-211
• /
• 2005

This paper makes an effort to compare the recently evolved soft classification method based on Linear Spectral Mixture Modeling (LSMM) with the traditional hard classification methods based on Iterative Self-Organizing Data Analysis (ISODATA) and Maximum Likelihood Classification (MLC) algorithms in order to achieve appropriate results for mapping, monitoring and preserving valuable coastal wetland ecosystems of southern India using Indian Remote Sensing Satellite (IRS) 1C/1D LISS-III and Landsat-5 Thematic Mapper image data. ISODATA and MLC methods were attempted on these satellite image data to produce maps of 5, 10, 15 and 20 wetland classes for each of three contrast coastal wetland sites, Pitchavaram, Vedaranniyam and Rameswaram. The accuracy of the derived classes was assessed with the simplest descriptive statistic technique called overall accuracy and a discrete multivariate technique called KAPPA accuracy. ISODATA classification resulted in maps with poor accuracy compared to MLC classification that produced maps with improved accuracy. However, there was a systematic decrease in overall accuracy and KAPPA accuracy, when more number of classes was derived from IRS-1C/1D and Landsat-5 TM imagery by ISODATA and MLC. There were two principal factors for the decreased classification accuracy, namely spectral overlapping/confusion and inadequate spatial resolution of the sensors. Compared to the former, the limited instantaneous field of view (IFOV) of these sensors caused occurrence of number of mixture pixels (mixels) in the image and its effect on the classification process was a major problem to deriving accurate wetland cover types, in spite of the increasing spatial resolution of new generation Earth Observation Sensors (EOS). In order to improve the classification accuracy, a soft classification method based on Linear Spectral Mixture Modeling (LSMM) was described to calculate the spectral mixture and classify IRS-1C/1D LISS-III and Landsat-5 TM Imagery. This method considered number of reflectance end-members that form the scene spectra, followed by the determination of their nature and finally the decomposition of the spectra into their endmembers. To evaluate the LSMM areal estimates, resulted fractional end-members were compared with normalized difference vegetation index (NDVI), ground truth data, as well as those estimates derived from the traditional hard classifier (MLC). The findings revealed that NDVI values and vegetation fractions were positively correlated ($r^2$= 0.96, 0.95 and 0.92 for Rameswaram, Vedaranniyam and Pitchavaram respectively) and NDVI and soil fraction values were negatively correlated ($r^2$ =0.53, 0.39 and 0.13), indicating the reliability of the sub-pixel classification. Comparing with ground truth data, the precision of LSMM for deriving moisture fraction was 92% and 96% for soil fraction. The LSMM in general would seem well suited to locating small wetland habitats which occurred as sub-pixel inclusions, and to representing continuous gradations between different habitat types.

• Korean Journal of Remote Sensing
• /
• v.22 no.6
• /
• pp.553-563
• /
• 2006

This study aims to produce land-cover maps of Korean peninsula using multi-temporal MODIS (Moderate Resolution Imaging Spectroradiometer) imagery. To solve the low spatial resolution of MODIS data and enhance classification accuracy, Linear Spectral Mixture Analysis (LSMA) was employed. LSMA allowed to determine the fraction of each surface type in a pixel and develop vegetation, soil and water fraction images. To eliminate clouds, MVC (Maximum Value Composite) was utilized for vegetation fraction and MinVC (Minimum Value Composite) for soil fraction image respectively. With these images, using ISODATA unsupervised classifier, southern part of Korean peninsula was classified to low and mid level land-cover classes. The results showed that vegetation and soil fraction images reflected phenological characteristics of Korean peninsula. Paddy fields and forest could be easily detected in spring and summer data of the entire peninsula and arable land in North Korea. Secondly, in low level land-cover classification, overall accuracy was 79.94% and Kappa value was 0.70. Classification accuracy of forest (88.12%) and paddy field (85.45%) was higher than that of barren land (60.71%) and grassland (57.14%). In midlevel classification, forest class was sub-divided into deciduous and conifers and field class was sub-divided into paddy and field classes. In mid level, overall accuracy was 82.02% and Kappa value was 0.6986. Classification accuracy of deciduous (86.96%) and paddy (85.38%) were higher than that of conifers (62.50%) and field (77.08%).

• Journal of the Korea Academia-Industrial cooperation Society
• /
• v.22 no.5
• /
• pp.7-14
• /
• 2021

Tomato crops are highly affected by tomato diseases, and if not prevented, a disease can cause severe losses for the agricultural economy. Therefore, there is a need for a system that quickly and accurately diagnoses various tomato diseases. In this paper, we propose a system that classifies nine diseases as well as healthy tomato plants by applying various pretrained deep learning-based CNN models trained on an ImageNet dataset. The tomato leaf image dataset obtained from PlantVillage is provided as input to ResNet, Xception, and DenseNet, which have deep learning-based CNN architectures. The proposed models were constructed by adding a top-level classifier to the basic CNN model, and they were trained by applying a 5-fold cross-validation strategy. All three of the proposed models were trained in two stages: transfer learning (which freezes the layers of the basic CNN model and then trains only the top-level classifiers), and fine-tuned learning (which sets the learning rate to a very small number and trains after unfreezing basic CNN layers). SGD, RMSprop, and Adam were applied as optimization algorithms. The experimental results show that the DenseNet CNN model to which the RMSprop algorithm was applied output the best results, with 98.63% accuracy.

• Journal of Digital Convergence
• /
• v.17 no.2
• /
• pp.163-170
• /
• 2019

Over the past decade, the development of the Web explosively increased the data. Feature selection step is an important step in extracting valuable data from a large amount of data. This study proposes a novel opinion mining model based on combining feature selection (FS) methods with Word embedding to vector (Word2vec) and BOW (Bag-of-words). FS methods adopted for this study are CFS (Correlation based FS) and IG (Information Gain). To select an optimal FS method, a number of classifiers ranging from LR (logistic regression), NN (neural network), NBN (naive Bayesian network) to RF (random forest), RS (random subspace), ST (stacking). Empirical results with electronics and kitchen datasets showed that LR and ST classifiers combined with IG applied to BOW features yield best performance in opinion mining. Results with laptop and restaurant datasets revealed that the RF classifier using IG applied to Word2vec features represents best performance in opinion mining.

• Journal of Internet Computing and Services
• /
• v.22 no.1
• /
• pp.23-30
• /
• 2021

This paper proposes a noise-tolerant image classification system using multiple autoencoders. The development of deep learning technology has dramatically improved the performance of image classifiers. However, if the images are contaminated by noise, the performance degrades rapidly. Noise added to the image is inevitably generated in the process of obtaining and transmitting the image. Therefore, in order to use the classifier in a real environment, we have to deal with the noise. On the other hand, the autoencoder is an artificial neural network model that is trained to have similar input and output values. If the input data is similar to the training data, the error between the input data and output data of the autoencoder will be small. However, if the input data is not similar to the training data, the error will be large. The proposed system uses the relationship between the input data and the output data of the autoencoder, and it has two phases to classify the images. In the first phase, the classes with the highest likelihood of classification are selected and subject to the procedure again in the second phase. For the performance analysis of the proposed system, classification accuracy was tested on a Gaussian noise-contaminated MNIST dataset. As a result of the experiment, it was confirmed that the proposed system in the noisy environment has higher accuracy than the CNN-based classification technique.