• KIPS Transactions on Software and Data Engineering
• /
• v.11 no.1
• /
• pp.29-34
• /
• 2022

We propose a novel network architecture and build dataset for recognizing clickable objects on mobile device screens. The data was collected based on clickable objects on the mobile device screen that have numerous resolution, and a total of 24,937 annotation data were subdivided into seven categories: text, edit text, image, button, region, status bar, and navigation bar. We use the Deconvolution Single Shot Detector as a baseline, the backbone network with Squeeze-and-Excitation blocks, the Single Shot Detector layer structure to derive inference results and the Feature pyramid networks structure. Also we efficiently extract features by changing the input resolution of the existing 1:1 ratio of the network to a 1:2 ratio similar to the mobile device screen. As a result of experimenting with the dataset we have built, the mean average precision was improved by up to 101% compared to baseline.

• Journal of Information Processing Systems
• /
• v.16 no.6
• /
• pp.1424-1436
• /
• 2020

Although huge progress has been made in current image segmentation work, there are still no efficient segmentation strategies for tree image which is taken from natural environment and contains complex background. To improve those problems, we propose a method for tree image segmentation combining adaptive mean shifting with image abstraction. Our approach perform better than others because it focuses mainly on the background of image and characteristics of the tree itself. First, we abstract the original tree image using bilateral filtering and image pyramid from multiple perspectives, which can reduce the influence of the background and tree canopy gaps on clustering. Spatial location and gray scale features are obtained by step detection and the insertion rule method, respectively. Bandwidths calculated by spatial location and gray scale features are then used to determine the size of the Gaussian kernel function and in the mean shift clustering. Furthermore, the flood fill method is employed to fill the results of clustering and highlight the region of interest. To prove the effectiveness of tree image abstractions on image clustering, we compared different abstraction levels and achieved the optimal clustering results. For our algorithm, the average segmentation accuracy (SA), over-segmentation rate (OR), and under-segmentation rate (UR) of the crown are 91.21%, 3.54%, and 9.85%, respectively. The average values of the trunk are 92.78%, 8.16%, and 7.93%, respectively. Comparing the results of our method experimentally with other popular tree image segmentation methods, our segmentation method get rid of human interaction and shows higher SA. Meanwhile, this work shows a promising application prospect on visual reconstruction and factors measurement of tree.

• The KIPS Transactions:PartB
• /
• v.11B no.3
• /
• pp.327-336
• /
• 2004

For visual inspection system in real industrial environment, it is one of most important tasks to design fast and stable pattern matching algorithm. This paper presents an edge-based point correlation algorithm avoiding full search in visual inspection system. Conventional algorithms based on NGC(normalized gray-level correlation) have to overcome some difficulties for applying to automated inspection system in factory environment. First of all, NGC algorithms need high time complexity and thus high performance hardware to satisfy real-time process. In addition, lighting condition in realistic factory environments if not stable and therefore intensity variation from uncontrolled lights gives many roubles for applying directly NGC as pattern matching algorithm in this paper, we propose an algorithm to solve these problems from using thinned and binarized edge data and skipping full point search with edge-map analysis. A point correlation algorithm with the thinned edges is introduced with image pyramid technique to reduce the time complexity. Matching edges instead of using original gray-level pixel data overcomes NGC problems and pyramid of edges also provides fast and stable processing. All proposed methods are preyed from experiments using real images.

• Journal of the Korean Society of Radiology
• /
• v.13 no.5
• /
• pp.749-755
• /
• 2019

In this study, we investigated the effects of tube voltage (kVp) and tube current (mAs) on an exposure index (EI) applied to a digital radiography. There used an inverter type digital X-ray generator and an image receptor (IR) utilized a portable wireless detector without a grid. Radiographic images were acquired using a cone pyramid phantom produced using a 3D printer. The X-ray tube voltage was increased from 40 kVp to 120 kVp with 10 kVp increment, and the tube current was increased from 1 mAs to 128 mAs with the twice increment. As a result, kVp had logarithmic relationship with the EI as high $R^2$ value, and mAs had a very high linear relationship too. Also, there was a high correlation between the area dose and the EI of the IR, with the $R^2$ value being 0.76 or more. In conclusion, it had showed that mAs affected the EI linearly, and that it could be advantageous to adjust the easy-to-predict mAs to maintain proper image qualities.

• Journal of the Korea Computer Graphics Society
• /
• v.4 no.2
• /
• pp.33-45
• /
• 1998

볼륨 가시화(volume visualization)는 3차원 볼륨 데이터로부터 의미 있는 가시적 정보를 추출하도록 도와주는 연구분야이다. 볼륨 렌더링(volume rendering)은 볼륨 데이터로부터 영상(image)을 얻는 기술을 말하는데, 이러한 렌더링 기법 중 물체공간(object space)에 기반한 스플래팅(splatting) 기법은 볼륨 데이터에 존재하는 응집성(coherence)의 이점을 이용할 수 있는 팔진트리(octree)나 피라미드(pyramid)와 같은 계층구조를 적용하기 쉽다. 본 논문에서는 볼륨 데이터에 팔진트리를 적용한 기존의 스플래팅 기법에 영상공간(image space)에서의 응집성의 이점을 이용하기 위한 계층구조로 4진트리(quadtree)와 범위트리(range tree)를 적용하는 새로운 스플래팅 기법을 제안한다. 이 기법은 볼륨 데이터내의 불투명한 복셀(voxel)들에 의해 가려지는 복셀들에 대한 방문을 가능한 한 피함으로써 전체적인 스플래팅의 속도를 향상시킨다. 이 기법은 잘 알려진 팔진트리, 4진트리 그리고 범위트리를 사용함으로써 그 구현이 쉽고, 추가적으로 많은 메모리를 사용하지 않으면서도 렌더링의 속도를 효율적으로 향상시킬 수 있는 기법이다.

• Proceedings of the IEEK Conference
• /
• 2003.11a
• /
• pp.237-240
• /
• 2003

We implement a system for 3D structure reconstruction from multiple 2D images. It uses geometric primitives such as box, wedge, pyramid, etc, each having translation, rotation, and scale parameters. Primitives are marked on input images with GUI (Graphic User Interface). Lines made by projection of primitives onto an image correspond to marked line segments of the image. Error function is defined by disparity between them and is minimized by downhill simplex method. By assigning relationship between models, the number of parameters to solve can be decreased and the resultant models become more accurate To share variables among other models also reduces computational complexity. Experiments using real images have shown that the proposed method successfully reconstructs 3D structure.

• Journal of the Semiconductor & Display Technology
• /
• v.6 no.3
• /
• pp.55-59
• /
• 2007

Wavelet Transform is amenable to efficient algorithms. So wavelet transform was adopted many signal processing and communication applications. For example, the wavelet transform was adopted as the transform for JPEG2000. However, Wavelet has weakness about smoothness along the contours and limited directional information. Hence, recently, some new transforms have been introduced to take advantage of this property. So we use to other transform, called contourlet transform in compression. In this paper, we propose a new method for image compression based on the contourlet transform, which has been recently introduced. Contourlet transform has a good result about images with smooth contours. Moreover, Contourlet is feasible multiresolution and multidirection expansion using non-separable filter bank. This treatise shows a good image representation after compressing using contourlet transform.

• The Journal of Korea Institute of Information, Electronics, and Communication Technology
• /
• v.2 no.3
• /
• pp.39-52
• /
• 2009

Scale Invariant Feature Transform (SIFT) is an effective algorithm in object recognition, panorama stitching, and image matching, however, due to its complexity, real time processing is difficult to achieve with software approaches. This paper proposes using a reconfigurable hardware processor with integer half kernel. The integer half kernel Gaussian reduces the Gaussian pyramid complexity in about half [] and the reconfigurable processor carries out a parallel implementation of a full search Fast SIFT algorithm. We use a low memory, fine grain single instruction stream multiple data stream (SIMD) pixel processor that is currently being developed. This implementation fully exposes the available parallelism of the SIFT algorithm process and exploits the processing and I/O capabilities of the processor which results in a system that can perform real time image and video compression. We apply this novel implementation to images and measure the effectiveness. Experimental simulation results indicate that the proposed implementation is capable of real time applications.

• JSTS:Journal of Semiconductor Technology and Science
• /
• v.12 no.2
• /
• pp.150-161
• /
• 2012

Eye detection is widely used in applications, such as face recognition, driver behavior analysis, and human-computer interaction. However, it is difficult to achieve real-time performance with software-based eye detection in an embedded environment. In this paper, we propose a parallel hardware architecture for real-time eye detection. We use the AdaBoost algorithm with modified census transform(MCT) to detect eyes on a face image. We parallelize part of the algorithm to speed up processing. Several downscaled pyramid images of the eye candidate region are generated in parallel using the input face image. We can detect the left and the right eye simultaneously using these downscaled images. The sequential data processing bottleneck caused by repetitive operation is removed by employing a pipelined parallel architecture. The proposed architecture is designed using Verilog HDL and implemented on a Virtex-5 FPGA for prototyping and evaluation. The proposed system can detect eyes within 0.15 ms in a VGA image.

• Journal of Digital Contents Society
• /
• v.7 no.1
• /
• pp.53-58
• /
• 2006

In this paper, We proposed extraction method using EZW a specific character and then add watermark significant coefficient of image. After wavelet transform in image, the significant coefficient value add to watermark information image. In this method, the locations of nonzero wavelet coefficients are encoded with a tree structure, called zerotree, which can exploit the self-similarity of the pyramid decomposition across different scales. The simulation shows that this method provides a superior performance over conventional method and can be successfully applied to the application areas that requires of progressive transmission and search for image data.