• Journal of the Korea Institute of Information and Communication Engineering
• /
• v.13 no.3
• /
• pp.594-601
• /
• 2009

Shooting simulation systems not only reduce a great amount of expense and time for military exercises but also prevent accidents. In particular, the shooting simulation systems using laser beam have an advantage which is very similar to the shooting exercise that uses real bullets. However, real time technique for laser beam recognition in a target image is necessary. The method proposed in this paper takes a difference image from two adjacent image frames. Then a thresholding is applied on this difference image to discriminate laser beam from background. To decide the threshold value the intensity distribution of background points is modeled assuming normal distribution. Then a noise reduction and a region segmentation are applied on the binary image to find the position of a laser beam. The time complexity of this process depends on the size of an image multiplied by the size of a mask used in the noise reduction process. The experimental result showed that the accuracy of the system was 93.3%. Even in the inaccurate cases the beam was always found in the resultant region.

• JSTS:Journal of Semiconductor Technology and Science
• /
• v.13 no.2
• /
• pp.157-169
• /
• 2013

Scale Invariant Feature Transform (SIFT) generates image features widely used to match objects in different images. Previous work on hardware-based SIFT implementation requires excessive internal memory and hardware logic [1]. In this paper, a new hardware organization is proposed to implement SIFT with less memory and hardware cost than the previous work. To this end, a parallel Gaussian filter bank is adopted to eliminate the buffers that store intermediate results because parallel operations allow all intermediate results available at the same time. Furthermore, the processing order is changed from the raster-scan order to the block-by-block order so that the line buffer size storing the source image is also reduced. These techniques trade the reduction of memory size with a slight increase of the execution time and external memory bandwidth. As a result, the memory size is reduced by 94.4%. The proposed hardware for SIFT implementation includes the Descriptor generation block, which is omitted in the previous work [1]. The addition of the hardwired descriptor generation improves the computation speed by about 30 times when compared with the previous work.

• Korean Journal of Remote Sensing
• /
• v.35 no.6_1
• /
• pp.959-971
• /
• 2019

CNN (Convolutional Neural Network) is one representative deep learning algorithm, which can extract high-level spatial and spectral features, and has been applied for hyperspectral image classification. However, one significant drawback behind the application of CNNs in hyperspectral images is the high dimensionality of the data, which increases the training time and processing complexity. To address this problem, several CNN based hyperspectral image classification studies have exploited PCA (Principal Component Analysis) for dimensionality reduction. One limitation to this is that the spectral information of the original image can be lost through PCA. Although it is clear that the use of PCA affects the accuracy and the CNN training time, the impact of PCA for CNN based hyperspectral image classification has been understudied. The purpose of this study is to analyze the quantitative effect of PCA in CNN for hyperspectral image classification. The hyperspectral images were first transformed through PCA and applied into the CNN model by varying the size of the reduced dimensionality. In addition, 2D-CNN and 3D-CNN frameworks were applied to analyze the sensitivity of the PCA with respect to the convolution kernel in the model. Experimental results were evaluated based on classification accuracy, learning time, variance ratio, and training process. The size of the reduced dimensionality was the most efficient when the explained variance ratio recorded 99.7%~99.8%. Since the 3D kernel had higher classification accuracy in the original-CNN than the PCA-CNN in comparison to the 2D-CNN, the results revealed that the dimensionality reduction was relatively less effective in 3D kernel.

• Journal of Korea Society of Digital Industry and Information Management
• /
• v.7 no.4
• /
• pp.133-139
• /
• 2011

In this paper, we propose a hybrid filter for noise reduction. The proposed method adjusts rational filtering direction according to an edge in the image using median filtered data. Rational filter modulates the coefficients of a linear lowpass filter to limit its action in presence of image details. By the ratio of polynomials in the input variables, rational filter reduces noise adaptively. Median filter is widely used to reduce impulse noise, but removes some details for highly corrupted images. Also, desirable details are removed when the window size is large. Our proposed algorithm combines rational filter and median filter. Thus, proposed method not only preserves edge, but also reduces noise in uniform region. Experimental results show that our proposed method has better quality than those by existing median and rational filtering methods.

• Journal of Multimedia Information System
• /
• v.5 no.2
• /
• pp.131-138
• /
• 2018

This work describes a binary image based fast Difference of Gaussian (DoG) filter using zero-dimensional (0-d) convolution and state machine look up tables (LUTs) for image and video stitching hardware platforms. The proposed approach for using binary images to obtain DoG filtering can significantly reduce the data size compared to conventional gray scale based DoG filters, yet binary images still preserve the key features of the image such as contours, edges, and corners. Furthermore, the binary image based DoG filtering can be realized with zero-dimensional convolution and state machine LUTs which eliminates the major portion of the adder and multiplier blocks that are generally used in conventional DoG filter hardware engines. This enables fast computation time along with the data size reduction which can lead to compact and low power image and video stitching hardware blocks. The proposed DoG filter using binary images has been implemented with a FPGA (Altera DE2-115), and the results have been verified.

• 제어로봇시스템학회:학술대회논문집
• /
• 2005.06a
• /
• pp.1244-1248
• /
• 2005

In this paper, we propose fuzzy c-means (FCM) to solve recognition errors in invariant range image, multi-pose face recognition. Scale, center and pose error problems were solved using geometric transformation. Range image face data was digitized into range image data by using the laser range finder that does not depend on the ambient light source. Then, the digitized range image face data is used as a model to generate multi-pose data. Each pose data size was reduced by linear reduction into the database. The reduced range image face data was transformed to the gradient face model for facial feature image extraction and also for matching using the fuzzy membership adjusted by fuzzy c-means. The proposed method was tested using facial range images from 40 people with normal facial expressions. The output of the detection and recognition system has to be accurate to about 93 percent. Simultaneously, the system must be robust enough to overcome typical image-acquisition problems such as noise, vertical rotated face and range resolution.

• Proceedings of the Korean Society of Broadcast Engineers Conference
• /
• 2012.11a
• /
• pp.52-53
• /
• 2012

Belief propagation (BP) is a commonly used global energy minimization algorithm for solving stereo matching problem in 3D reconstruction. However, it requires large memory bandwidth and data size. In this paper, we propose a novel memory-efficient algorithm of BP in stereo matching on the Graphics Processing Units (GPU). The data size and transfer bandwidth are significantly reduced by storing only a part of the whole message. In order to maintain the accuracy of the matching result, the local messages are reconstructed using shared memory available in GPU. Experimental result shows that there is almost an order of reduction in the global memory consumption, and 21 to 46% saving in memory bandwidth when compared to the conventional algorithm. The implementation result on a recent GPU shows that we can obtain 22.8 times speedup in execution time compared to the execution on CPU.

• Journal of the Korean Society of Radiology
• /
• v.15 no.1
• /
• pp.21-28
• /
• 2021

Computed tomography (CT) has the problem of having more radiation exposure compared to other radiographic apparatus. There is a low-dose imaging technique for reducing exposure, but it has a disadvantage of increasing noise in the image. To compensate for this, various noise reduction algorithms have been developed that improve image quality while reducing the exposure dose of patients, of which the median modified Wiener filter (MMWF) algorithm that can be effectively applied to CT devices with excellent time resolution has been presented. The purpose of this study is to optimize the mask size of MMWF algorithm and to see the excellence of noise reduction of MMWF algorithm for existing algorithms. After applying the MMWF algorithm with each mask sizes set from the MASH phantom abdominal images acquired using the MATLAB program, which includes Gaussian noise added, and compared the values of root mean square error (RMSE), peak signal-to-noise ratio (PSNR), coefficient correlation (CC), and universal image quality index (UQI). The results showed that RMSE value was the lowest and PSNR, CC and UQI values were the highest in the 5 x 5 mask size. In addition, comparing Gaussian filter, median filter, Wiener filter, and MMWF with RMSE, PSNR, CC, and UQI by applying the optimized mask size. As a result, the most improved RMSE, PSNR, CC, and UQI values were showed in MMWF algorithms.

• Journal of Korea Multimedia Society
• /
• v.13 no.6
• /
• pp.817-824
• /
• 2010

Linear Discriminant Analysis (LDA) has been successfully applied for dimension reduction in face recognition. However, LDA requires the transformation of a face image to a one-dimensional vector and this process can cause the correlation information among neighboring pixels to be disregarded. On the other hand, 2D-LDA uses 2D images directly without a transformation process and it has been shown to be superior to the traditional LDA. Nevertheless, there are some problems in 2D-LDA. First, it is difficult to determine the optimal number of feature vectors in a reduced dimensional space. Second, the size of rectangular windows used in 2D-LDA makes strong impacts on classification accuracies but there is no reliable way to determine an optimal window size. In this paper, we propose a new algorithm to overcome those problems in 2D-LDA. We adopt an ensemble approach which combines several classifiers obtained by utilizing various window sizes. And a practical method to determine the number of feature vectors is also presented. Experimental results demonstrate that the proposed method can overcome the difficulties with choosing an optimal window size and the number of feature vectors.

• Journal of the Korean Wood Science and Technology
• /
• v.37 no.6
• /
• pp.556-565
• /
• 2009

Automated visual grading system of lumber needs correct input image. In order to create a correct image of domestic red pine lumber 3.6 m long feeding on a conveyer, part images were captured using area sensor and template matching algorithm was applied to merge part images. Two kinds of template matching algorithms and six kinds of template sizes were adopted in this operation. Feature extracted method appeared to have more excellent image merging performance than fixed template method. Error length was attributed to a decline of similarity related by difference of partial brightness on a part image, specific pattern and template size. The mismatch part was repetitively generated at the long grain. The best size of template for image merging was $100{\times}100$ pixels. In a further study, assignment of exact template size, preprocessing of image merging for reduction of brightness difference will be needed to improve image merging.