• Journal of Korea Multimedia Society
• /
• v.23 no.7
• /
• pp.803-811
• /
• 2020

This paper proposes an image classification algorithm that transforms the number of convolution layers in the residual block of ResNet, CNN's representative method. The proposed method modified the structure of 34/50 layer of ResNet structure. First, we analyzed the performance of small and many convolution layers for the structure consisting of only shortcut and 3 × 3 convolution layers for 34 and 50 layers. And then the performance was analyzed in the case of small and many cases of convolutional layers for the bottleneck structure of 50 layers. By applying the results, the best classification method in the residual block was applied to construct a 34-layer simple structure and a 50-layer bottleneck image classification model. To evaluate the performance of the proposed image classification model, the results were analyzed by applying to the cifar10 dataset. The proposed 34-layer simple structure and 50-layer bottleneck showed improved performance over the ResNet-110 and Densnet-40 models.

• ETRI Journal
• /
• v.43 no.2
• /
• pp.272-287
• /
• 2021

This article presents a multi-dimensional spatial pattern analysis of crime events in San Francisco. Our analysis includes the impact of spatial resolution on hotspot identification, temporal effects in crime spatial patterns, and relationships between various crime categories. In this work, crime prediction is viewed as a classification problem. When predictions for a particular category are made, a binary classification-based model is framed, and when all categories are considered for analysis, a multiclass model is formulated. The proposed crime-prediction model (HotBlock) utilizes spatiotemporal analysis for predicting crime in a fixed spatial region over a period of time. It is robust under variation of model parameters. HotBlock's results are compared with baseline real-world crime datasets. It is found that the proposed model outperforms the standard DeepCrime model in most cases.

• Journal of the Korean Institute of Telematics and Electronics B
• /
• v.30B no.12
• /
• pp.67-72
• /
• 1993

An adaptive block truncation coding(BTC) using human visual system(HVS) is proposed. To reduce visible blocking effect at sensitive area in HVS. a new category classification coefficient is proposed. The categroy classification coefficient was derived by combining the modified HVS and standard deviation. By computer simulations, we showed that the proposed method reduced blocking effect at low bit rate coding more than the conventional Hui's method.

• Proceedings of the Korean Society of Broadcast Engineers Conference
• /
• 1998.06a
• /
• pp.33-36
• /
• 1998

Most image restoration problems are ill-posed and need to e regularized. A difficult task in image regularization is to avoid smoothing of image edges. In this paper, were proposed an edge-preserving image restoration algorithm using block-based edge classification. In order to exploit the local image characteristics, we classify image blocks into edge and no-edge blocks. We then apply an adaptive constrained least squares (CLS) algorithm to eliminate noise around the edges. Experimental results demonstrate that the proposed algorithm can preserve image edges during the regularization process.

• Journal of the Korea Institute of Information and Communication Engineering
• /
• v.16 no.1
• /
• pp.59-64
• /
• 2012

This paper addresses a blind estimation and classification algorithm of the speech compression methods by using analysis on the structure of compressed bitstreams. Various speech compression methods including vocoders are developed in order to transmit or store the speech signals at very low bitrates. As a key feature, the vocoders contain the block structure inevitably. In classification of each compression method, we use the Measure of Inter-Block Correlation (MIBC) to check whether the bitstream includes the block structure or not, and to estimate the block length. Moreover, for the compression methods with the same block length, the proposed algorithm estimates the corresponding compression method correctly by using that each compression method has different correlation characteristics in each bit location. Experimental results indicate that the proposed algorithm classifies the speech compression methods robustly for various types and lengths of speech signals in noisy environment.

• Journal of the Korean Institute of Intelligent Systems
• /
• v.25 no.5
• /
• pp.483-488
• /
• 2015

This paper presents a local binary pattern(LBP) for effectively classifying textures, which is based on the multidimensional intensity difference between the adjacent pixels in the block image. The intensity difference by considering the a extent of 4 directional changes(verticality, horizontality, diagonality, inverse diagonality) in brightness between the adjacent pixels is applied to reduce the computation load as a results of decreasing the levels of histogram for classifying textures of image. And the binary patterns that is represented by the relevant intensities within a block image, is also used to effectively classify the textures by accurately reflecting the local attributes. The proposed method has been applied to classify 24 block images from USC Texture Mosaic #2 of 128*128 pixels gray image. The block images are different in size and texture. The experimental results show that the proposed method has a speedy classification and makes a free size block images classify possible. In particular, the proposed method gives better results than the conventional LBP by increasing the range of histogram level reduction as the block size becomes larger.

• Journal of Service Research and Studies
• /
• v.5 no.2
• /
• pp.103-110
• /
• 2015

Recently, Classification of characters in a 3D FPS game has emerged as a very significant issue. In this study, We propose the game character Classification method using Image Region Segmentation of the extracting meaningful object in a simple operation. In this method, first used a non-linear RGB color model and octree color quantization scheme. The input image represented a less than 20 quantized color and uses a small number of meaningful color histogram. And then, the image divided into small blocks, calculate the degree of similarity between the color histogram intersection and adjacent block in block units. Because, except for the block boundary according to the texture and to extract only the boundaries of the object block. Set a region by these boundary blocks as a game object and can be used for FPS game play. Through experiment, we obtain accuracy of more than 80% for Classification method using each feature. Thus, using this property, characters could be classified effectively and it draws the game more speed and strategic actions as a result.

• Journal of the Institute of Electronics Engineers of Korea SP
• /
• v.39 no.2
• /
• pp.87-95
• /
• 2002

In this paper, we proposed a fast multi-resolution motion estimation(MRME) algorithm. This algorithm exploits the half-pixel accuracy motion estimation(HPAME) for exact motion vectors in the baseband and block classification for the reduction of bit amounts and computational loads. Generally, as the motion vector in the baseband are used as initial motion vector in the high frequency subbands, it has crucial effect on quality of the motion compensated image. For this reason, we exploit HPAME in the motion estimation for the baseband. But HPAME requires additional bit and computational loads so that we use block classification for the selective motion estimation in the high frequency subbands to compensate these problems. In result, we could reduce the bit rate and computational load at the similar image quality with conventional MRME. The superiority of the proposed algorithm was confirmed by the computer simulation.

• The Journal of the Institute of Internet, Broadcasting and Communication
• /
• v.21 no.6
• /
• pp.155-161
• /
• 2021

Gesture recognition analytics through a camera in real time have been widely studied in recent years. Since a small number of features from human joints are extracted, low accuracy of classifying models is get in conventional gesture recognition studies. In this paper, CBAM (Convolutional Block Attention Module) with high accuracy for classifying images is proposed as a classification model and algorithm calculating the angle of joints depending on actions is presented to solve the issues. Employing five exercise gestures images from the fitness posture images provided by AI Hub, the images are applied to the classification model. Important 8-joint angles information for classifying the exercise gestures is extracted from the images by using MediaPipe, a graph-based framework provided by Google. Setting the features as input of the classification model, the classification model is learned. From the simulation results, it is confirmed that the exercise gestures are classified with high accuracy in the proposed model.

• Proceedings of the IEEK Conference
• /
• 1998.10a
• /
• pp.907-910
• /
• 1998

In this paper, we present a comparative performance analysis of several blocking artifact reduction algorithms. For the performance analysis, we propose a block boundary region classification algorithm which classifies each horizontal and vertical block boundary into four regions using brightness change near the block boundary. The PSNR performance of each algorithm is compared. The MSE according to each block boundary region is also compared. Experimental results show that the wavelet transform based blocking artifact reduction algorithms have better performance over the other methods.