• Journal of Korea Game Society
• /
• v.12 no.5
• /
• pp.67-78
• /
• 2012

Summed area table allows filtering of arbitrary-width box regions for every pixel in constant time per pixel. This characteristic makes it beneficial in image processing applications where the sum or average of the surrounding pixel intensity is required. Although calculating the summed area table of an image data is primarily a memory bound job consisting of row or column-wise summation, previous works had to endure excessive access to the high latency global memory in order to exploit data parallelism. In this paper, we propose an efficient algorithm for generating the summed area table in the GPGPU environment where the input is decomposed into square sub-images with intermediate data that are propagated between them. By doing so, the global memory access is almost halved compared to the previous methods making an efficient use of the available memory bandwidth. The results show a substantial increase in performance.

• KIPS Transactions on Software and Data Engineering
• /
• v.2 no.11
• /
• pp.809-814
• /
• 2013

Summed area table (SAT) is a data structure in which the sum of pixel values in an arbitrary rectangular area can be represented by the linear combination of four pixel values. Since SAT serially accumulates the pixel values from an image corner to the other corner, a high-resolution image can yield overflow in a floating-point representation. In this paper, we present a new SAT construction technique, which accumulates only the residuals from the linearly-regressed representation of an image and thereby significantly reduces the accumulation errors. Also, we propose a method to find the integral of the linear regression in constant time using double integral. We performed experiments on the image reconstruction, and the results showed that our approach more reduces the accumulation errors than the conventional fixed-offset SAT.

• Proceedings of the Korea Information Processing Society Conference
• /
• 2013.05a
• /
• pp.386-388
• /
• 2013

합산 영역 테이블(Summed Area Table)을 사용하면 현재 픽셀 주변으로 임의의 사각 영역의 평균을 모든 픽셀을 읽을 필요 없이, 단 4번의 픽셀의 합과 차로 표시할 수 있다. 그러나 많은 픽셀의 값이 누적되는 경우 부동소수점 표현의 정밀도가 떨어지는 문제가 발생한다. 따라서 본 논문에서는 합산 영역 테이블의 정밀도를 향상시키기 위한 방법으로 선형 회귀분석(linear regression)을 이용한 오프셋을 사용할 것을 제안한다. 회귀분석을 통해 구축한 다항식을 통해 픽셀 그리고 채널 별로 다른 오프셋을 적용하여 정밀도를 효과적으로 향상하였다.

• Journal of Korea Multimedia Society
• /
• v.9 no.10
• /
• pp.1266-1275
• /
• 2006

In this paper, we propose a real-lime face detection and tracking algorithm using AdaBoost(Adaptive Boosting) algorithm. The proposed algorithm consists of two levels such as the face detection and the face tracking. First, the face detection used the eight-wavelet feature models which ate very simple. Each feature model applied to variable size and position, and then create initial feature set. The intial feature set and the training images which were consisted of face images, non-face images used the AdaBoost algorithm. The basic principal of the AdaBoost algorithm is to create final strong classifier joining linearly weak classifiers. In the training of the AdaBoost algorithm, we propose SAT(Summed-Area Table) method. Face tracking becomes accomplished at real-time using the position information and the size information of detected face, and it is extended view region dynamically using the fan-Tilt camera. We are setting to move center of the detected face to center of the Image. The experiment results were amply satisfied with the computational efficiency and the detection rates. In real-time application using Pan-Tilt camera, the detecter runs at about 12 frames per second.

• Journal of Institute of Control, Robotics and Systems
• /
• v.20 no.9
• /
• pp.930-935
• /
• 2014

A Gabor filter is a linear filter used for edge detectionas frequency and orientation representations of Gabor filters are similar to those of the human visual system. In this thesis, we propose a pedestrian detection algorithm using a Gabor filter bank. In order to extract the features of the pedestrian, we use various image processing algorithms and data structure algorithms. First, color image segmentation is performed to consider the information of the RGB color space. Second, histogram equalization is performed to enhance the brightness of the input images. Third, convolution is performed between a Gabor filter bank and the enhanced images. Fourth, statistical values are calculated by using the integral image (summed area table) method. The calculated statistical values are used for the feature matrix of the pedestrian area. To evaluate the proposed algorithm, the INRIA pedestrian database and SVM (Support Vector Machine) are used, and we compare the proposed algorithm and the HOG (Histogram of Oriented Gradient) pedestrian detector, presentlyreferred to as the methodology of pedestrian detection algorithm. The experimental results show that the proposed algorithm is more accurate compared to the HOG pedestrian detector.

• Journal of Intelligence and Information Systems
• /
• v.20 no.2
• /
• pp.137-148
• /
• 2014

Recommender system has become one of the most important technologies in e-commerce in these days. The ultimate reason to shop online, for many consumers, is to reduce the efforts for information search and purchase. Recommender system is a key technology to serve these needs. Many of the past studies about recommender systems have been devoted to developing and improving recommendation algorithms and collaborative filtering (CF) is known to be the most successful one. Despite its success, however, CF has several shortcomings such as cold-start, sparsity, gray sheep problems. In order to be able to generate recommendations, ordinary CF algorithms require evaluations or preference information directly from users. For new users who do not have any evaluations or preference information, therefore, CF cannot come up with recommendations (Cold-star problem). As the numbers of products and customers increase, the scale of the data increases exponentially and most of the data cells are empty. This sparse dataset makes computation for recommendation extremely hard (Sparsity problem). Since CF is based on the assumption that there are groups of users sharing common preferences or tastes, CF becomes inaccurate if there are many users with rare and unique tastes (Gray sheep problem). This study proposes a new algorithm that utilizes Social Network Analysis (SNA) techniques to resolve the gray sheep problem. We utilize 'degree centrality' in SNA to identify users with unique preferences (gray sheep). Degree centrality in SNA refers to the number of direct links to and from a node. In a network of users who are connected through common preferences or tastes, those with unique tastes have fewer links to other users (nodes) and they are isolated from other users. Therefore, gray sheep can be identified by calculating degree centrality of each node. We divide the dataset into two, gray sheep and others, based on the degree centrality of the users. Then, different similarity measures and recommendation methods are applied to these two datasets. More detail algorithm is as follows: Step 1: Convert the initial data which is a two-mode network (user to item) into an one-mode network (user to user). Step 2: Calculate degree centrality of each node and separate those nodes having degree centrality values lower than the pre-set threshold. The threshold value is determined by simulations such that the accuracy of CF for the remaining dataset is maximized. Step 3: Ordinary CF algorithm is applied to the remaining dataset. Step 4: Since the separated dataset consist of users with unique tastes, an ordinary CF algorithm cannot generate recommendations for them. A 'popular item' method is used to generate recommendations for these users. The F measures of the two datasets are weighted by the numbers of nodes and summed to be used as the final performance metric. In order to test performance improvement by this new algorithm, an empirical study was conducted using a publically available dataset - the MovieLens data by GroupLens research team. We used 100,000 evaluations by 943 users on 1,682 movies. The proposed algorithm was compared with an ordinary CF algorithm utilizing 'Best-N-neighbors' and 'Cosine' similarity method. The empirical results show that F measure was improved about 11% on average when the proposed algorithm was used