• KSII Transactions on Internet and Information Systems (TIIS)
• /
• v.15 no.1
• /
• pp.147-165
• /
• 2021

In this work, we empirically evaluated the efficiency of the recent EfficientNetB0 model to identify and diagnose malaria parasite infections in blood smears. The dataset used was collected and classified by relevant experts from the Lister Hill National Centre for Biomedical Communications (LHNCBC). We prepared our samples with minimal image transformations as opposed to others, as we focused more on the feature extraction capability of the EfficientNetB0 baseline model. We applied transfer learning to increase the initial feature sets and reduced the training time to train our model. We then fine-tuned it to work with our proposed layers and re-trained the entire model to learn from our prepared dataset. The highest overall accuracy attained from our evaluated results was 94.70% from fifty epochs and followed by 94.68% within just ten. Additional visualization and analysis using the Gradient-weighted Class Activation Mapping (Grad-CAM) algorithm visualized how effectively our fine-tuned EfficientNetB0 detected infections better than other recent state-of-the-art DCNN models. This study, therefore, concludes that when fine-tuned, the recent EfficientNetB0 will generate highly accurate deep learning solutions for the identification of malaria parasites in blood smears without the need for stringent pre-processing, optimization, or data augmentation of images.

• The Journal of Korean Institute of Information Technology
• /
• v.17 no.5
• /
• pp.77-82
• /
• 2019

Diabetic retinopathy is a complicated form of diabetes due to circulatory disorder in the peripheral blood vessels of the retina. We segment the microvessel for diagnosing diabetic retinophathy. The conventional methods using filter and features can segment the thick blood vessels, but it has relatively weak for segmenting fine blood vessels. In pre-processing step, noise reduction filter and histogram equalization are applied to suppress the noise and enhance the image contrast. Then, deep learning technique is used for pixel-by-pixel segmentation. The accuracy of conventional methods is between 90% to 94%, while the proposed method has improved as 95% accuracy. There is a problem of segmentation error around the optic disc and exudate due to the network depth. However the accuracy can be improved by modifying the network architecture in the future.

• Korean Journal of Remote Sensing
• /
• v.38 no.1
• /
• pp.45-56
• /
• 2022

The Ieodo Ocean Research Station (IORS) lies between the exclusive economic zone (EEZ) boundaries of Korea, Japan, and China. The geographical positioning of the IORS makes it ideal for monitoring ships in the area. In this study, we introduce ship monitoring results by Automatic Identification System (AIS) and the Broadband 3G^TM radar, which has been developed for use in small ships using the Frequency Modulated Continuous Wave (FMCW) technique. AIS and FMCW radar data were collected at IORS from November 23th to 30th, 2013. The acquired FMCW radar data was converted to 2-D binary image format over pre-processing, including the internal and external noise filtering. The ship positions detected by FMCW radar images were passed into a tracking algorithm. We then compared the detection and tracking results from FMCW radar with AIS information and found that they were relatively well matched. Tracking performance is especially good when ships are across from each other. The results also show good monitoring capability for small fishing ships, even those not equipped with AIS or with a dysfunctional AIS.

• Proceedings of the Korean Society of Computer Information Conference
• /
• 2021.01a
• /
• pp.45-46
• /
• 2021

의료영상이란 의료영상장비로부터 DICOM이라는 의료영상표준에 따라 저장되며, 의료영상관리 시스템인 PACS를 통해 관리된다. 이러한, 의료영상장비 ICT기술이 융합되어 급격하게 발전되고 있으며 다양한 의료영상장치가 개발되어지고 있다. 하지만, 기술력은 높아지고 있으나 개발된 의료영상장비로부터 촬영된 영상품질관리에 대한 문제점이 제기되고 있다. 이와 관련하여 다기관의 의료영상장비 개발과 해당 기기로부터 수집된 의료영상에 대한 품질을 관리할 필요성이 증가하고 있다. 따라서 코로나 19와 같은 상황에서 의료기기 개발 지원과 관리를 비대면 관리서비스 시스템 개발과 의료영상장치 개발 정도를 관리할 수 있을 뿐만 아니라 의료영상에 대한 품질까지 모니터링하여 및 개선 할 수 있는 시스템을 제안하고자 한다.

• Proceedings of the Korean Institute of Information and Commucation Sciences Conference
• /
• 2021.10a
• /
• pp.97-99
• /
• 2021

As the aging society continues in recent years, interest in dementia is increasing. Among them, Alzheimer's disease is a degenerative brain disease that accounts for the largest percentage of all dementia patients, with the medical community currently not offering clear prevention and treatment for Alzheimer's disease, and the importance of early treatment and early prevention is emphasized. In this paper, we intend to find the most efficient activation function by combining various activation functions centering on convolutional neural networks using MRI datasets of normal people and patients with Alzheimer's disease. In addition, it is intended to be used as a dementia classification modeling suitable for the medical field in the future through Alzheimer's dementia classification modeling.

• Journal of the Korea Society of Computer and Information
• /
• v.4 no.2
• /
• pp.120-134
• /
• 1999

Image processing steps consist of image acquisition, pre-processing, region segmentation and recognition, and the images are easily corrupted by noise during the data transmission, data capture, and data processing. Impulse noise and gaussian noise are major noises, which can occur during the process. Many filters such as mean filter, median filter, weighted median filter, Cheikh filter, and Kyu-cheol Lee filter were proposed as spatial noise reduction filters so far. Many researches have been focused on the reduction of impulse noise, but comparatively the research in the reduction of gaussian noise has been neglected. For the reduction of gaussian noise, subset averaged median filter, using median information and subset average information of pixels in a window. was proposed. At this time, consider of the window size as 3$^{*}$3 pixel. The window is divided to 4 subsets consisted of 4 pixels. First of all, we calculate the average value of each subset, and then find the median value by sorting the average values and center pixel's value. In this paper, a better reduction of gaussian noise was proved. The proposed algorithms were implemented by ANSI C language on a Sun Ultra 2 for testing purposes and the effects and results of the filter in the various levels of noise and images were proposed by comparing the values of PSNR, MSE, and RMSE with the value of the other existing filtering methods.thods.

• The Journal of Korea Institute of Information, Electronics, and Communication Technology
• /
• v.2 no.1
• /
• pp.29-35
• /
• 2009

In this paper by using the image taken from the rear of the vehicle to effectively extract the license plate and how to recognize the characters appearing in the offer. How to existing research on the entire video by following the pre-edge (edge) images to obtain yijinhwa. Qualified heopeu in a binary image (Hough) to convert the horizontal and vertical lines to obtain, using the characteristics of the plates to extract the license plate area. The problem with this method, the processing time is so difficult to handle real-time status of irregular points, and visual contrast with yagangwan border does not appear in the plates to extract the license plate area is that it is not. In addition, the rear of the vehicle license plate area from images taken using the characteristics of the plates myeongamgap changes sutjapok in the area, background area and the number number area of the region confirmed the contrast of the car and identified the number and the number of 42 of distance to extract the license plate area. How to research, the existing damage to the border of the plate to fail to extract the license plate area, a matter of hours to resolve problems in real-time, practical application is processed. Chapter 100 as the results of the experiment the sample video image in a car that far experiment results automatically read license plates have been able to extract the license plate and failing to represent 13% of images, character recognition result of failing to represent the image was 0.4%

• The Journal of Korean Institute of Communications and Information Sciences
• /
• v.30 no.4C
• /
• pp.215-222
• /
• 2005

In this paper, we propose a fingerprint verification algorithm for the embedded system based on the minutia extracted using the image quality, the minutia structure, and the Sequency and the orientation of ridges. After the pre- and the post-processing, the true minutia are selected, thus it shows high reliability in the fingerprint verification. In matching process, we consider the errors caused by shift, rotation, and pressure when acquiring the fingerprint image and reduce the matching time by applying a local matching instead of a full matching to select the reference pair. The proposed algorithm has been designed and verified in Arm920T environment and various techniques for the realtime process have been applied. Time taken from the fingerprint registration through out the matching is 0.541 second that is relevant for the realtime applications. The FRR (False Reject Rate) and FAR (False Accept Rate) show 0.079 and 0.00005 respectively.

• Journal of the Korean Society of Surveying, Geodesy, Photogrammetry and Cartography
• /
• v.26 no.2
• /
• pp.129-137
• /
• 2008

This study focuses on the radiometric normalization, which is one of the pre-processing steps to apply the change detection technique fur hyperspectral images. The PIFs which had radiometric consistency under the time interval were automatically extracted by applying spectral angle, and used as sample pixels for linear regression of the radiometric normalization. We also dealt with the problem about the number of PIFs for linear regression with iteratively quantitative methods. The results were assessed in comparison with image regression, histogram matching, and FLAASH. In conclusion, we show that linear regression method with PIFs can carry out the efficient result for radiometric normalization.

• Journal of the Korea Institute of Information and Communication Engineering
• /
• v.4 no.5
• /
• pp.1051-1066
• /
• 2000

According to the recent advances of digital encoding technologies and computing power, large amounts of multimedia informations such as image, graphic, audio and video are fully used in multimedia systems through Internet. By this, diverse retrieval mechanisms are required for users to search dedicated informations stored in multimedia systems, and especially it is preferred to use contents-based retrieval method rather than text-type keyword retrieval method. In this paper, we propose a new contents-based indexing and searching algorithm which aims to get both high efficiency and high retrieval performance. To achieve these objectives, firstly the proposed algorithm classifies images by a pre-processing process of edge extraction, range division, and multiple filtering, and secondly it searches the target images using spatial and textural characteristics of colors, which are extracted from the previous process, in a image. In addition, we describe the simulation results of search requests and retrieval outputs for several images of company's trade-mark using the proposed contents-based retrieval algorithm based on wavelet.