• Journal of KIISE:Software and Applications
• /
• v.32 no.2
• /
• pp.85-98
• /
• 2005

Recently, a lot of researches about mobile vision using Personal Digital Assistant(PDA) has been attempted. Many CPUs for PDA are integer CPUs, which have no floating-computation component. It results in slow computation of the algorithms peformed by vision system or image processing, which have much floating-computation. In this paper, in order to resolve this weakness, we propose the Client(PDA)/server(PC) architecture which is connected to each other with a wireless LAN, and we construct the system with pipelining processing using two CPUs of the Client(PDA) and the Server(PC) in image sequence. The Client(PDA) extracts tentative text regions using Edge Density(ED). The Server(PC) uses both the Multi-1.aver Perceptron(MLP)-based texture classifier and Connected Component(CC)-based filtering for a definite text extraction based on the Client(PDA)'s tentativel99-y extracted results. The proposed method leads to not only efficient text extraction by using both the MLP and the CC, but also fast running time using Client(PDA)/server(PC) architecture with the pipelining processing.

• The Journal of Korean Institute of Communications and Information Sciences
• /
• v.29 no.10C
• /
• pp.1425-1432
• /
• 2004

Because of the high performance with the edge regions, the existing LSE(Least Square Estimation) method provides much better results than other methods. However, since it emphasizes not oがy edge components but also noise components, some part of interpolated images looks like unnatural. It also requires very high computational complexity and memory for implementation. We propose a new LSE interpolation method which requires much lower complexity and memory, but provides better performance than the existing method. To reduce the computational complexity, we propose and adopt a simple sample window and a direction detector to reduce the size of memory without blurring image. To prevent from emphasizing noise components, the hi-linear interpolation method is added in the LSE formula. The simulation results show that the proposed method provides better subjective and objective performance with love. complexity than the existing method.

• Transactions on Semiconductor Engineering
• /
• v.2 no.1
• /
• pp.17-20
• /
• 2024

This paper presents a novel lightweight object detection model tailored for low-powered edge devices, addressing the limitations of traditional resource-intensive computer vision models. Our proposed detector, inspired by the Single Shot Detector (SSD), employs a compact yet robust network design. Crucially, it integrates an 'enhancer block' that significantly boosts its efficiency in detecting smaller objects. The model comprises two primary components: the Light_Block for efficient feature extraction using Depth-wise and Pointwise Convolution layers, and the Enhancer_Block for enhanced detection of tiny objects. Trained from scratch on the Udacity Annotated Dataset with image dimensions of 300x480, our model eschews the need for pre-trained classification weights. Weighing only 5.5MB with approximately 0.43M parameters, our detector achieved a mean average precision (mAP) of 27.7% and processed at 140 FPS, outperforming conventional models in both precision and efficiency. This research underscores the potential of lightweight designs in advancing object detection for edge devices without compromising accuracy.

• Journal of the Institute of Electronics Engineers of Korea SP
• /
• v.46 no.3
• /
• pp.105-111
• /
• 2009

In this paper, we propose a locally adaptive image restoration method based on nonlinear extrapolation in frequency domain. In general, the conventional method causes ringing artifacts on the object boundary. To solve this problem, we introduce an improved restoration method which considers textures of an image block. In the proposed method, a blurred image is divided into several blocks, and each block is classified into three groups; simple, one edge, and complex blocks according to the contained texture. Depending on the classification result, adaptive nonlinear extrapolation is applied to each block in a blurred image. Experimental results show that the proposed algorithm can achieve higher quality image in both subjective and objective views as compared with the conventional method.

• Journal of Digital Contents Society
• /
• v.16 no.4
• /
• pp.615-622
• /
• 2015

Recently virtual advertising is located in an important area of interest in the TV market by convenience of application and reduction of cost. The methods of inserting a virtual advertising in broadcasting are Up-link that method insert the image through the production equipment of the broadcasting station and dispatch equipment and technical personnel in the shooting and Down-streaming that method insert a virtual image automatically in relay video using image processing technology. In recent years, the image processing technology is an important research area in the virtual advertising area for automatically insertion of advertising images. In this paper, we propose the method to extract less-informative region in sports video using image processing. The proposed method extracts less-Informative region through rectangle detection of Hough transform and analysis of color histogram distribution.

• The KIPS Transactions:PartB
• /
• v.9B no.3
• /
• pp.327-336
• /
• 2002

Texture analysis is an important technique in many image understanding areas, such as perception of surface, object, shape and depth. But the previous works are intend to the issue of only texture segment, that is not capable of acquiring recognition information. No unsupervised method is basased on the recognition of texture in image. we propose a novel approach for efficient texture image analysis that uses unsupervised learning schemes for the texture recognition. The self-organization neural network for multiple texture image identification is based on block-based clustering and merging. The texture features used are the angle and magnitude in orientation-field that might be different from the sample textures. In order to show the performance of the proposed system, After we have attempted to build a various texture images. The final segmentation is achieved by using efficient edge detection algorithm applying to block-based dilation. The experimental results show that the performance of the system Is very successful.

• The KIPS Transactions:PartB
• /
• v.11B no.4
• /
• pp.509-516
• /
• 2004

Semantic interpretation of image is incomplete without some mechanism for understanding semantic content that is not directly visible. For this reason, human assisted content-annotation through natural language is an attachment of textual description to image. However, keyword-based retrieval is in the level of syntactic pattern matching. In other words, dissimilarity computation among terms is usually done by using string matching not concept matching. In this paper, we propose a method for computerized semantic similarity calculation In WordNet space. We consider the edge, depth, link type and density as well as existence of common ancestors. Also, we have introduced method that applied similarity measurement on semantic image retrieval. To combine wi#h the low level features, we use the spatial color distribution model. When tested on a image set of Microsoft's 'Design Gallery Line', proposed method outperforms other approach.

• Journal of the Korea Institute of Information and Communication Engineering
• /
• v.21 no.7
• /
• pp.1371-1378
• /
• 2017

In this paper, we present an upsampling technique for depth map image using selective bilateral weights and a color weight using laplacian function. These techniques prevent color texture copy problem, which problem appears in existing upsamplers uses bilateral weight. First, we construct a high-resolution image using the bicubic interpolation technique. Next, we detect a color texture region using pixel value differences of depth and color image. If an interpolated pixel belongs to the color texture edge region, we calculate weighting values of spatial and depth in $3{\times}3$ neighboring pixels and compute the cost value to determine the boundary pixel value. Otherwise we use color weight instead of depth weight. Finally, the pixel value having minimum cost is determined as the pixel value of the high-resolution depth image. Simulation results show that the proposed algorithm achieves good performance in terns of PSNR comparison and subjective visual quality.

• Computers and Concrete
• /
• v.10 no.2
• /
• pp.135-147
• /
• 2012

At mesoscale, concrete is considered as a three-phase composite material consisting of the aggregate particles, the cement matrix and the interfacial transition zone (ITZ). The reconstruction of the internal structures for concrete composites requires the identification of the boundary of the aggregate particles and the cement matrix using digital imaging technology followed by post-processing through MATLAB. A parameter study covers the subsection transformation, median filter, and open and close operation of the digital image sample to obtain the optimal parameter for performing the image processing technology. The subsection transformation is performed using a grey histogram of the digital image samples with a threshold value of [120, 210] followed by median filtering with a $16{\times}16$ square module based on the dimensions of the aggregate particles and their internal impurity. We then select a "disk" tectonic structure with a specific radius, which performs open and close operations on the images. The edges of the aggregate particles (similar to the original digital images) are obtained using the canny edge detection method. The finite element model at mesoscale can be established using the proposed image processing technology. The location of the crack determined through the numerical method is identical to the experimental result, and the load-displacement curve determined through the numerical method is in close agreement with the experimental results. Comparisons of the numerical and experimental results show that the proposed image processing technology is highly effective in reconstructing the internal structures of concrete composites.

• Journal of the Korea Institute of Information and Communication Engineering
• /
• v.14 no.6
• /
• pp.1469-1475
• /
• 2010

This paper presents a problem of existing encryption methods using pseudo-random numbers based on MLCA or complemented MLCA and proposes a method to resolve this problem. The existing encryption methods have a problem which the edge of original image appear on encrypted image because the image have color similarity of adjacent pixels. In this proposed method, we transform the value and spatial coordinates of all pixels by using pseudo-random numbers based on MLCA. This method can resolve the problem of existing methods and improve the level of encryption by encrypting pixel coordinates and pixel values of original image. The effectiveness of the proposed method is proved by conducting histogram and key space analysis.