• Journal of the Semiconductor & Display Technology
• /
• v.6 no.2 s.19
• /
• pp.49-52
• /
• 2007

Characteristics of an active pixel switch readout circuit were studied by SPICE simulation. A simple readout circuit consists of an operation amplifier, a diode, and a down-counter was suggested, and its successful operation was verified by showing that the differences in the detected signal intensity are accordingly converted to modulation of the voltage pulses generated by the comparator. A scheme to use these pulses to generate the original image was also put forward.

• Korean Journal of Remote Sensing
• /
• v.25 no.1
• /
• pp.21-30
• /
• 2009

The purposes of this paper are to present a selection method of neighboring pixels whose local statistics are similar to the center pixel and combine the selection result with mean curvature diffusion filter to reduce noises in remote sensed imagery. The order of selection of neighboring pixels is critical, especially for finding a pixel belonging to the homogeneous region, since the statistics of the homogeneous region vary according to the selection order. An effective strategy for selecting neighboring pixels, which uses rank-order differences vector obtained by computing the intensity differences between the center pixel and neighboring pixels and arranging them in ascending order, is proposed in this paper. By using region growing method, we divide the elements of the rank-ordered differences vector into two groups, homogeneous rank-ordered differences vector and outlier rank-ordered differences vector. The mean curvature diffusion filter is combined with a line process, which chooses selectively diffusion coefficient of the neighboring pixels belonging into homogeneous rank-ordered differences vector. Experimental results using an aerial image and a TerraSAR-X satellite image showed that the proposed method reduced more efficiently noises than some conventional adaptive filters using all neighboring pixels in updating the center pixel.

• Journal of KIISE:Software and Applications
• /
• v.33 no.4
• /
• pp.411-417
• /
• 2006

Motion Estimation (ME) is an important part of video coding process and it takes the largest amount of computation in video compression. Half-pixel and quarter-pixel motion estimation can improve the video compression rate at the cost of higher computational complexity In this paper, we suggest a new efficient low-complexity algorithm for half-pixel and quarter pixel motion estimation. It is based on the experimental results that the sum of absolute differences(SAD) shows parabolic shape and thus can be approximated by using interpolation techniques. The sub-pixel motion vector is searched from the minimum SAD integer-pixel motion vector. The sub-pixel search direction is determined toward the neighboring pixel with the lowest SAD among 8 neighbors. Experimental results show that more than 20% reduction in computation time can be achieved without affecting the quality of video.

• The Journal of Korean Institute of Communications and Information Sciences
• /
• v.29 no.6C
• /
• pp.848-856
• /
• 2004

In this paper, a fast stereo matching algorithm based on pixel-wise matching strategy, which can get a stable and accurate disparity map, is proposed. Since a stereo image pair has small differences each other and the differences between left and right images are just caused by horizontal shifts with some order, the matching using a large window will not be needed within a given search range. However, disparity results of conventional pixel-based matching methods are somewhat unstable and wrinkled, the principal direction of disparities is checked by the accumulated cost along a path on array with the dynamic programming method. Experimental results showed that the proposed method could remove almost all disparity noise and set a good quality disparity map in very short time.

• The Transactions of the Korean Institute of Electrical Engineers D
• /
• v.54 no.3
• /
• pp.148-155
• /
• 2005

A seam-line determination algorithm is proposed to determine image border-line in mosaicing using the transformation of gray value differences and dynamic programming. Since visually good border-line is the one along which pixel differences are as small as possible, it can be determined in association with an optimal path finding algorithm. A well-known effective optimal path finding algorithm is the Dynamic Programming (DP). Direct application of the dynamic programming to the seam-line determination causes the distance effect, in which seam-line is affected by its length as well as the gray value difference. In this paper, an adaptive cost transform algorithm with which the distance effect is suppressed is proposed in order to utilize the dynamic programming on the transformed pixel difference space. Also, a figure of merit which is the summation of fixed number of the biggest pixel difference on the seam-line (SFBPD) is suggested as an evaluation measure of seamlines. The performance of the proposed algorithm has been tested in both quantitively and visually on various kinds of images.

• Annual Conference of KIPS
• /
• 2010.11a
• /
• pp.625-628
• /
• 2010

This paper proposes a novel feature extraction method for unsupervised multispectral image segmentation based in one dimensional combined neighborhood differences (1D CND). In contrast with the original CND, which is applied with traditional image, 1D CND is computed on a single pixel with various bands. The proposed algorithm utilizes the sign of differences between bands of the pixel. The difference values are thresholded to form a binary codeword. A binomial factor is assigned to these codeword to form another unique value. These values are then grouped to construct the 1D CND feature image where is used in the unsupervised image segmentation. Various experiments using two LANDSAT multispectral images have been performed to evaluate the segmentation and classification accuracy of the proposed method. The result shows that 1D CND feature outperforms the spectral feature, with average classification accuracy of 87.55% whereas that of spectral feature is 55.81%.

• Convergence Security Journal
• /
• v.13 no.3
• /
• pp.9-15
• /
• 2013

Calculating of matching cost is an important for efficient stereo matching. To investigate the performance of matching process, the concepts of the existing methods are introduced. Also we analyze the performance and merits of them. The simplest matching costs assume constant intensities at matching image locations. We consider matching cost functions which can be distinguished between pixel-based and window-based approaches. The Pixel-based approach includes absolute differences (AD) and sampling-intensitive absolute differences (BT). The window-based approach includes the sum of the absolute differences, the sum of squared differences, the normalized cross-correlation, zero-mean normalized cross-correlation, census transform, and the absolute differences census transform (AD-Census). We evaluate matching cost functions in terms of accuracy and time complexity. In terms of the accuracy, AD-Census method shows the lowest matching error ratio (the best solution). The ZNCC method shows the lowest matching error ratio in non-occlusion and all evaluation part. But it performs high matching error ratio at the discontinuities evaluation part due to blurring effect in the boundary. The pixel-based AD method shows a low complexity in terms of time complexity.

• Journal of Institute of Control, Robotics and Systems
• /
• v.12 no.3
• /
• pp.248-254
• /
• 2006

It is necessary to improve the exactness and adaptation of the working environment for the intelligent robot system. The vision sensor have been studied for a long time at this points. However, it has many processes and difficulties for the real usages. This paper proposes a visual servoing in the virtual environment to support OLP(Off-Line-Programming) path compensation and supplement the problem of complexity of the old kinematical calibration. Initial robot path could be compensated by pixel differences between real and virtual image. This method removes the varies calibrations and 3D reconstruction process in real working space. To show the validity of the proposed approach, virtual space servoing with stereo camera is carried out with WTK and openGL library for a KUKA-6R manipulator and updated real robot path.

• Journal of Korean Academy of Oral and Maxillofacial Radiology
• /
• v.29 no.2
• /
• pp.397-405
• /
• 1999

Purpose: To clarify the usefulness and the limitation of Digora system/sup (R)/ by evaluating the physical characteristics as the quantitative image on Image Plate(Ip). Materials and Methods: Radiograms were taken by Heliodent MD(Siemens Co.. Germany) with the image plate for adult. Cu-step wedge as reference material. and three pieces of dry mandibular bone. Image analysis was performed by single color enhancement. density measurement with histogram. The relationship between the exposure conditions and the distribution of the pixel values of the image. the variation of pixel values of each step of Cu-step wedge at two different area and Cu-equivalent value of three pieces of dry mandibular bone measure by the conversion equation. Results: There was no linear relationship between the exposure condition and the average pixel value of the image. of which the distribution was not even. The pixel value differences between the center portion and the periphery were ranged from 60 to 70 in vertical plane and from 15 to 26 in horizontal plane. Two plot profile formed at two different areas of the Cu-step wedge were different. The measured Cu-equivalent values showed the discrepancy among the times of measurement. Conclusion: As above results. Image Plate(Ip) of Digora system/sup (R)/ showed the limitation as the quantitative image. The physical property of IP was expected to need to be compensated for the quantitative evaluation of the bone or others

• Journal of Information Technology Services
• /
• v.15 no.4
• /
• pp.85-95
• /
• 2016

Recently, deep learning and machine learning have attracted considerable attention and many supporting frameworks appeared. In artificial intelligence field, a large body of research is underway to apply the relevant knowledge for complex problem-solving, necessitating the application of various learning algorithms and training methods to artificial intelligence systems. In addition, there is a dearth of performance evaluation of decision making agents. The decision making agent that can find optimal solutions by using reinforcement learning methods designed through this research can collect raw pixel data observed from dynamic environments and make decisions by itself based on the data. The decision making agent uses convolutional neural networks to classify situations it confronts, and the data observed from the environment undergoes preprocessing before being used. This research represents how the convolutional neural networks and the decision making agent are configured, analyzes learning performance through a value-based algorithm and a policy-based algorithm : a Deep Q-Networks and a Policy Gradient, sets forth their differences and demonstrates how the convolutional neural networks affect entire learning performance when using pixel data. This research is expected to contribute to the improvement of artificial intelligence systems which can efficiently find optimal solutions by using features extracted from raw pixel data.