• Journal of Satellite, Information and Communications
• /
• v.7 no.2
• /
• pp.97-103
• /
• 2012

A very low complexity QRD-M algorithm based on limited search area is proposed for MIMO systems. The conventional QRD-M algorithm calculates Euclidean distance between all constellation symbols and the temporary detection symbol at each layer. We found that performance will not be degraded even if we adaptively restrict the search area of the candidate symbols only to the neighboring points of temporary detection symbol according to the channel condition at each layer. As a channel condition indicator, we employ the channel gain ratio among the layers without necessity of SNR estimation. The simulation results show that the proposed scheme effectively achieves near optimal performance while maintaining the overall average computation complexity much smaller than the conventional QRD-M algorithm.

• KIPS Transactions on Software and Data Engineering
• /
• v.3 no.2
• /
• pp.93-100
• /
• 2014

3D object recognition using only 2D images is a difficult work because each images are generated different to according to the view direction of cameras. Because SIFT algorithm defines the local features of the projected images, recognition result is particularly limited in case of input images with strong perspective transformation. In this paper, we propose the object recognition method that improves SIFT algorithm by using several sequential images captured from rotating 3D object around a rotation axis. We use the geometric relationship between adjacent images and merge several images into a generated feature space during recognizing object. To clarify effectiveness of the proposed algorithm, we keep constantly the camera position and illumination conditions. This method can recognize the appearance of 3D objects that previous approach can not recognize with usually SIFT algorithm.

• Journal of Broadcast Engineering
• /
• v.24 no.2
• /
• pp.217-226
• /
• 2019

In this paper, we propose a filtering method that can improve the results of inverse tone-mapping using guided image filter. Inverse tone-mapping techniques have been proposed that convert LDR images to HDR. Recently, many algorithms have been studied to convert single LDR images into HDR images using CNN. Among them, there exists an algorithm for restoring pixel information using CNN which learned to restore saturated region. The algorithm does not suppress the noise in the non-saturation region and cannot restore the detail in the saturated region. The proposed algorithm suppresses the noise in the non-saturated region and restores the detail of the saturated region using a WGIF in the input image, and then applies it to the CNN to improve the quality of the final image. The proposed algorithm shows a higher quantitative image quality index than the existing algorithms when the HDR quantitative image quality index was measured.

• Journal of the Korean Institute of Telematics and Electronics
• /
• v.27 no.2
• /
• pp.65-72
• /
• 1990

The potential of neural networks for the recognition of the printed Korean characters is examined. In spite of good classification capability of neural networks, it is difficult to train a neural network to recognize Korean characters. The difficulty is due to a large number of Korean characters, the similarities among the characters, and the large number of data from the character images. To reduce the input image data, DC components are extracted from each input images. These preprocessed data are used as input to the neural network. The output nodes are composed to represent the characteristics of Korean characters. A MLP (multilayer perceptron) with one hidden layer was trained with a modified BEP algorithm, This method gives good recognition rate for the standard positioned characters of more than 2,300. The result shows that neural networks are well suited for the recognition of printed Korean characters.

• Journal of the Korean Institute of Telematics and Electronics
• /
• v.27 no.2
• /
• pp.144-150
• /
• 1990

This paper describes radix-4 Fast Fourier Transform (FFT) Processor designed with the new twice perfect shuffle developed from a perfect shuffle used in radix-2 FFT algorithm. The FFT Processor consists of a butterfly arithmetic circuit, address generators for input, output and coefficient, input and output registers and controller. Also, it requires the external ROM for storage of coefficient and RAM for input and output. The butterfly circuit includes 12 bit-serial ($16{\times}8$) multipliers, adders, subtractors and delay shift registers. Operating on 25 MHz two phase clock, this processor can compute 256 point FFT in 6168 clocks, i.e. 247 us and provides flexibility by allowing the user to select any size among 4,16,64,and256points. Being fabricated with 2-um double metal CMOS process, it includes about 28000 transistors and 55 pads in $8.0{\times}8.2mm^2$area.

• Journal of the Korea Academia-Industrial cooperation Society
• /
• v.22 no.5
• /
• pp.251-256
• /
• 2021

Unlike general cameras, a high-speed camera capable of capturing a large number of frames per second can enable the advancement of some image processing technologies that have been limited so far. This paper proposes a method of removing undesirable noise from an high-speed input color image, and then detecting a human face from the noise-free image. In this paper, noise pixels included in the ultrafast input image are first removed by applying a bidirectional filter. Then, using RetinaFace, a region representing the person's personal information is robustly detected from the image where noise was removed. The experimental results show that the described algorithm removes noise from the input image and then robustly detects a human face using the generated model. The model-based face-detection method presented in this paper is expected to be used as basic technology for many practical application fields related to image processing and pattern recognition, such as indoor and outdoor building monitoring, door opening and closing management, and mobile biometric authentication.

• The Journal of Korean Institute of Communications and Information Sciences
• /
• v.29 no.8C
• /
• pp.1159-1174
• /
• 2004

In this paper, we present an efficient algorithm for region analysis of business card images acquired in a PDA by using DCT and information pixel density. The proposed method consists of three parts: region segmentation, information region classification, and text region classification. In the region segmentation, an input business card image is partitioned into 8 f8 blocks and the blocks are classified into information and background blocks using the normalized DCT energy in their low frequency bands. The input image is then segmented into information and background regions by region labeling on the classified blocks. In the information region classification, each information region is classified into picture region or text region by using a ratio of the DCT energy of horizontal and vertical edge components to that in low frequency band and a density of information pixels, that are black pixels in its binarized region. In the text region classification, each text region is classified into large character region or small character region by using the density of information pixels and an averaged horizontal and vertical run-lengths of information pixels. Experimental results show that the proposed method yields good performance of region segmentation, information region classification, and text region classification for test images of several types of business cards acquired by a PDA under various surrounding conditions. In addition, the error rates of the proposed region segmentation are about 2.2-10.1% lower than those of the conventional region segmentation methods. It is also shown that the error rates of the proposed information region classification is about 1.7% lower than that of the conventional information region classification method.

• Journal of the Korea Academia-Industrial cooperation Society
• /
• v.17 no.10
• /
• pp.61-67
• /
• 2016

With the current high distribution rate of smart devices and the recent development of computing technology, user interest in multimedia, such as photos, videos, and so on, has rapidly increased, which is a departure from the simple pattern of information retrieval. Because of these increasing interests, image processing techniques, which generate and process images for diverse applications, are being developed. In entertainment recently, there are some techniques that present a celebrity's image as a mosaic comprising many small images. In addition, studies into the mosaic technique are actively conducted. However, conventional mosaic techniques result in a long processing time as the number of database images increases, because they compare the images in the databases sequentially. Therefore, to increase search efficiency, this paper proposes an algorithm to generate a mosaic image using a discrete cosine transform (DCT) hash. The proposed photo mosaic-generation algorithm is composed of database creation and mosaic image generation. In database creation, it first segments images into blocks with a predefined size. And then, it computes and stores a DCT hash set for each segmented block. In mosaic generation, it efficiently searches for the most similar blocks in the database via DCT hash for every block of the input image, and then it generates the final mosaic. With diverse experimental results, the proposed photo mosaic-creation algorithm can effectively generate a mosaic, regardless of the various types of images and sizes.

• Smart Structures and Systems
• /
• v.29 no.1
• /
• pp.1-16
• /
• 2022

Despite recent breakthroughs in deep learning and computer vision fields, the pixel-wise identification of tiny objects in high-resolution images with complex disturbances remains challenging. This study proposes a modified U-net for tiny crack segmentation in real-world steel-box-girder bridges. The modified U-net adopts the common U-net framework and a novel Self-Attention-Self-Adaption (SASA) neuron as the fundamental computing element. The Self-Attention module applies softmax and gate operations to obtain the attention vector. It enables the neuron to focus on the most significant receptive fields when processing large-scale feature maps. The Self-Adaption module consists of a multiplayer perceptron subnet and achieves deeper feature extraction inside a single neuron. For data augmentation, a grid-based crack random elastic deformation (CRED) algorithm is designed to enrich the diversities and irregular shapes of distributed cracks. Grid-based uniform control nodes are first set on both input images and binary labels, random offsets are then employed on these control nodes, and bilinear interpolation is performed for the rest pixels. The proposed SASA neuron and CRED algorithm are simultaneously deployed to train the modified U-net. 200 raw images with a high resolution of 4928 × 3264 are collected, 160 for training and the rest 40 for the test. 512 × 512 patches are generated from the original images by a sliding window with an overlap of 256 as inputs. Results show that the average IoU between the recognized and ground-truth cracks reaches 0.409, which is 29.8% higher than the regular U-net. A five-fold cross-validation study is performed to verify that the proposed method is robust to different training and test images. Ablation experiments further demonstrate the effectiveness of the proposed SASA neuron and CRED algorithm. Promotions of the average IoU individually utilizing the SASA and CRED module add up to the final promotion of the full model, indicating that the SASA and CRED modules contribute to the different stages of model and data in the training process.

• KIPS Transactions on Software and Data Engineering
• /
• v.6 no.6
• /
• pp.315-320
• /
• 2017

Estimating the correct pose of augmented objects on the real camera view efficiently is one of the most important questions in image tracking area. In computer vision, Homography is used for camera pose estimation in augmented reality system with markerless. To estimating Homography, several algorithm like SURF features which extracted from images are used. Based on extracted features, Homography is estimated. For this purpose, RANSAC algorithm is well used to estimate homography and DCS-RANSAC algorithm is researched which apply constraints dynamically based on Constraint Satisfaction Problem to improve performance. In DCS-RANSAC, however, the dataset is based on pattern of feature distribution of images manually, so this algorithm cannot classify the input image, pattern of feature distribution is not recognized in DCS-RANSAC algorithm, which lead to reduce it's performance. To improve this problem, we suggest the KCS-RANSAC algorithm using K-means clustering in CS-RANSAC to cluster the images automatically based on pattern of feature distribution and apply constraints to each image groups. The suggested algorithm cluster the images automatically and apply the constraints to each clustered image groups. The experiment result shows that our KCS-RANSAC algorithm outperformed the DCS-RANSAC algorithm in terms of speed, accuracy, and inlier rate.