• Journal of the Korea Institute of Information and Communication Engineering
• /
• v.11 no.3
• /
• pp.608-618
• /
• 2007

Image binarization is an important preprocessing to identify the object of interest by dividing pixels into the background and object. We proposes efficient binarization method and a stroke reconstruction method of the low quality character image for an effective character recognition. First, the character image is binarized by using the both advantages of local and global thresholding method and then the noise elimination around the character stroke and the hole filling on the stoke by the analysis of the binarized stroke image are performed to enhance the quality of the character stroke. Proposed binarization algorithm for character image achieved an efficiency of both processing speed and performance by the adaptive threshold selection. Moreover, We could get a high qualify binary image by a stroke reconstruction of the step-by-step denoising process.

• Journal of the Korea Institute of Information and Communication Engineering
• /
• v.26 no.10
• /
• pp.1483-1489
• /
• 2022

Recently, due to the influence of the 4th industrial revolution and the development of communication media, various digital video equipment are being used in industrial fields. Image data is easily damaged by noise in the process of acquiring and transmitting and receiving from the camera and sensor, and since the damaged image has a great effect on the processing of the system, noise removal is essential. In this paper, a weight filter algorithm using a weight graph is proposed to restoration images damaged by high-density impulse noise. The proposed algorithm obtains a weight graph using pixel values inside the filtering mask of the image, and restores the image by applying the final weight to the filtering mask. Simulation was conducted to analyze the noise removal performance of the proposed algorithm, and the magnified image and PSNR were used to compare with the existing method. The resulting image of the proposed algorithm showed excellent performance by removing high-density impulse noise.

• Journal of Convergence for Information Technology
• /
• v.11 no.2
• /
• pp.41-46
• /
• 2021

The method of the edge detection algorithm based on grayscale mathematical morphology has the advantage that image noise can be removed and processed in parallel, and the operation speed is fast. However, the method of detecting the edge of an image using a single structural scale element may be affected by image information. The characteristics of grayscale morphology may be limited to the edge information result of the operation result by repeatedly performing expansion, erosion, opening, and containment operations by repeating structural elements. In this paper, we propose an edge detection algorithm that applies a structural element with strong directionality to noise and then applies weighted entropy to each pixel information in the element. The result of applying the multi-scale structural element applied to the image and the result of applying the directional weighted entropy were compared and analyzed, and the simulation result showed that the proposed algorithm is superior in edge detection.

• Journal of the Korean Society of Radiology
• /
• v.83 no.2
• /
• pp.344-359
• /
• 2022

Purpose To develop a denoising convolutional neural network-based image processing technique and investigate its efficacy in diagnosing breast cancer using low-dose mammography imaging. Materials and Methods A total of 6 breast radiologists were included in this prospective study. All radiologists independently evaluated low-dose images for lesion detection and rated them for diagnostic quality using a qualitative scale. After application of the denoising network, the same radiologists evaluated lesion detectability and image quality. For clinical application, a consensus on lesion type and localization on preoperative mammographic examinations of breast cancer patients was reached after discussion. Thereafter, coded low-dose, reconstructed full-dose, and full-dose images were presented and assessed in a random order. Results Lesions on 40% reconstructed full-dose images were better perceived when compared with low-dose images of mastectomy specimens as a reference. In clinical application, as compared to 40% reconstructed images, higher values were given on full-dose images for resolution (p < 0.001); diagnostic quality for calcifications (p < 0.001); and for masses, asymmetry, or architectural distortion (p = 0.037). The 40% reconstructed images showed comparable values to 100% full-dose images for overall quality (p = 0.547), lesion visibility (p = 0.120), and contrast (p = 0.083), without significant differences. Conclusion Effective denoising and image reconstruction processing techniques can enable breast cancer diagnosis with substantial radiation dose reduction.

• Journal of radiological science and technology
• /
• v.40 no.4
• /
• pp.589-594
• /
• 2017

The objective of this study was to evaluate the enhancement effects of the quantum denoising system (QDS) on brain CT images. This retrospective study was conducted with 45 adults who visited G Radiology located in Gyungbuk for having brain CT tests between Jul 2017 and Oct 2017 after receiving consents. Subjects were divided into a control group (A group; no QDS(-) application during the brain CT test) and a treatment group (B Group; QDS(+) application during the brain CT test). The following conclusions were obtained from the study. The noise values at the Pons part and the Vermis part were significantly (p<0.05) lower in B Group ($Pons=5.41{\pm}1.05HU$; $Vermis=5.28{\pm}0.73HU$) than A Group ($Pons=6.92{\pm}0.98HU$; Vermis=6.72). The SNR values at the Pons part and the Vermis part were significantly (p<0.05) higher in B Group ($Pons=7.28{\pm}2.56$; $Vermis=8.63{\pm}3.04$) than A Group ($Pons=5.21{\pm}1.28$; $Vermis=6.23{\pm}1.49$). In conclusion, the results of this study suggested that the application of QDS to the brain CT test would enhance the signal to noise ratio (SNR) and the contrast to noise ratio (CNR) to provide an image more appropriate for diagnosis.

• The Korean Journal of Applied Statistics
• /
• v.23 no.3
• /
• pp.595-607
• /
• 2010

This paper addresses the problem of reducing the speckle noise in SAR images by wavelet transformation, using a non-local means(NLM) filter originated for Gaussian noise removal. Log-transformed SAR image makes multiplicative speckle noise additive. Thus, non-local means filtering and wavelet thresholding are used to reduce the additive noise, followed by an exponential transformation. NLM filter is an image denoising method that replaces each pixel by a weighted average of all the similarly pixels in the image. But the NLM filter takes an acceptable amount of time to perform the process for all possible pairs of pixels. This paper, also proposes an alternative strategy that uses the t-test more efficiently to eliminate pixel pairs that are dissimilar. Extensive simulations showed that the proposed filter outperforms many existing filters terms of quantitative measures such as PSNR and DSSIM as well as qualitative judgments of image quality and the computational time required to restore images.

• Journal of the Korea Institute of Information and Communication Engineering
• /
• v.17 no.10
• /
• pp.2447-2453
• /
• 2013

The digital images are being degraded by noise in the process of acquisition, storage and transmission, Gaussian or impulse noise is the representative noise. Meanwhile, the image has lots of tendency to be degraded by complex noise, so various researches are being conducted for reducing these complex noise. In this paper, to remove complex noise, the algorithm processed by modified switching median filter and modified adaptive weighted filter according to the result after judging the kinds of noise is proposed. In the simulation result, excellent denoising capabilities. Furthermore, we compared proposed algorithm with existing methods for objective judgement, and PSNR(peak signal to noise ratio) is used by the criterion of judgement.

• Journal of Korea Multimedia Society
• /
• v.23 no.8
• /
• pp.965-976
• /
• 2020

We present in this work a segmentation method of E. coli bacterial images generated via phase contrast microscopy using a deep learning based hybrid feature generation. Unlike conventional machine learning methods that use the hand-crafted features, we adopt the denoising autoencoder in order to generate a precise and accurate representation of the pixels. We first construct a hybrid vector that combines original image, difference of Gaussians and image gradients. The created hybrid features are then given to a deep autoencoder that learns the pixels' internal dependencies and the cells' shape and boundary information. The latent representations learned by the autoencoder are used as the inputs of a softmax classification layer and the direct outputs from the classifier represent the coarse segmentation mask. Finally, the classifier's outputs are used as prior information for a graph partitioning based fine segmentation. We demonstrate that the proposed hybrid vector representation manages to preserve the global shape and boundary information of the cells, allowing to retrieve the majority of the cellular patterns without the need of any post-processing.

• Journal of Korea Society of Digital Industry and Information Management
• /
• v.8 no.1
• /
• pp.171-181
• /
• 2012

This paper introduces the double-density discrete wavelet transform(DWT) using quincunx sampling, which is a DWT that combines the double-density DWT and quincunx sampling method, each of which has its own characteristics and advantages. The double-density DWT is an improvement upon the critically sampled DWT with important additional properties: Firstly, It employs one scaling function and two distinct wavelets, which are designed to be offset from one another by one half. Secondly, the double-density DWT is overcomplete by a factor of two, and Finally, it is nearly shift-invariant. In two dimensions, this transform outperforms the standard DWT in terms of denoising; however, there is room for improvement because not all of the wavelets are directional. That is, although the double-density DWT utilizes more wavelets, some lack a dominant spatial orientation, which prevents them from being able to isolate those directions. A solution to this problem is a quincunx sampling method. The quincunx lattice is a sampling method in image processing. It treats the different directions more homogeneously than the separable two dimensional schemes. Proposed wavelet transformation can generate sub-images of multiple degrees rotated versions. Therefore, This method services good performance in image processing fields.

• Journal of the Korea Institute of Information and Communication Engineering
• /
• v.17 no.7
• /
• pp.1725-1731
• /
• 2013

Recently, as industrial society is developed as high digital information age, image processing is used in a variety of areas. However, image deterioration is still occurred by several noises in the data processing, and different kinds of researches are in progress for noise removal. Therefore, this paper limited the modified median filter by using standard deviation to removes impulse noise. The proposed algorithms divided into noise detection and noise removal, the noise free pixel is maintained, the noise pixel filtered. And it was compared with the existing methods for its objective judgment, using PSNR(peak signal to noise ratio) as standard criteria.