• Journal of the Korea Society of Computer and Information
• /
• v.17 no.12
• /
• pp.71-82
• /
• 2012

In this paper, we propose a license plate detection method with improved Adaboost learning and MCT (Modified Census Transform). The MCT represents the local structure patterns as integer numbered feature values which has robustness to illumination change and memory efficiency. However, since these integer values are discrete, a lookup table is needed to design a weak classifier for Adaboost learning. Some previous research efforts have focused on minimization of exponential criterion for Adaboost optimization. In this paper, a method that uses MCT and improved Adaboost learning based on Newton's optimization to exponential criterion is proposed for license plate detection. Experimental results on license patch images and field images demonstrate that the proposed method yields higher performance of detection rates with low false positives than the conventional method using the original Adaboost learning.

• International Journal of Fuzzy Logic and Intelligent Systems
• /
• v.12 no.2
• /
• pp.137-142
• /
• 2012

The temporal subtraction technique as one of computer aided diagnosis has been introduced in medical fields to enhance the interval changes such as formation of new lesions and changes in existing abnormalities on deference image. With the temporal subtraction technique radiologists can easily detect lung nodules on visual screening. Until now, two-dimensional temporal subtraction imaging technique has been introduced for the clinical test. We have developed new temporal subtraction method to remove the subtraction artifacts which is caused by mis-registration on temporal subtraction images of lungs on MDCT images. In this paper, we propose a new computer aided diagnosis scheme for automatic enhancing the lung nodules from the temporal subtraction of thoracic MDCT images. At first, the candidates regions included nodules are detected by the multiple threshold technique in terms of the pixel value on the temporal subtraction images. Then, a rule-base method and artificial neural networks is utilized to remove the false positives of nodule candidates which is obtained temporal subtraction images. We have applied our detection of lung nodules to 30 thoracic MDCT image sets including lung nodules. With the detection method, satisfactory experimental results are obtained. Some experimental results are shown with discussion.

• Biomolecules & Therapeutics
• /
• v.23 no.4
• /
• pp.379-385
• /
• 2015

The eye irritation potential of drug candidates or pharmaceutical ingredients should be evaluated if there is a possibility of ocular exposure. Traditionally, the ocular irritation has been evaluated by the rabbit Draize test. However, rabbit eyes are more sensitive to irritants than human eyes, therefore substantial level of false positives are unavoidable. To resolve this species difference, several three-dimensional human corneal epithelial (HCE) models have been developed as alternative eye irritation test methods. Recently, we introduced a new HCE model, MCTT HCE$^{TM}$ which is reconstructed with non-transformed human corneal cells from limbal tissues. Here, we examined if MCTT HCE$^{TM}$ can be employed to evaluate eye irritation potential of solid substances. Through optimization of washing method and exposure time, treatment time was established as 10 min and washing procedure was set up as 4 times of washing with 10 mL of PBS and shaking in 30 mL of PBS in a beaker. With the established eye irritation test protocol, 11 solid substances (5 non-irritants, 6 irritants) were evaluated which demonstrated an excellent predictive capacity (100% accuracy, 100% specificity and 100% sensitivity). We also compared the performance of our test method with rabbit Draize test results and in vitro cytotoxicity test with 2D human corneal epithelial cell lines.

• Journal of Microbiology and Biotechnology
• /
• v.29 no.3
• /
• pp.454-464
• /
• 2019

Salmonellosis is a highly contagious bacterial disease that threatens both human and poultry health. Tests that can detect Salmonella in the field are urgently required to facilitate disease control and for epidemiological investigations. Here, we combined loop-mediated isothermal amplification (LAMP) with a chromatographic lateral flow dipstick (LFD) to rapidly and accurately detect Salmonella. LAMP primers were designed to target the Salmonella invA gene. LAMP conditions were optimized by adjusting the ratio of inner to outer primers, $MgSO_4$ concentration, dNTP mix concentration, amplification temperature, and amplification time. We evaluated the specificity of our novel LAMP-LFD method using six Salmonella species and six related non-Salmonella strains. All six of the Salmonella strains, but none of the non-Salmonella strains, were amplified. LAMP-LFD was sensitive enough to detect concentrations of Salmonella enterica subsp. enterica serovar Pullorum genomic DNA as low as $89fg/{\mu}l$, which is 1,000 times more sensitive than conventional PCR. When artificially contaminated feed samples were analyzed, LAMP-LFD was also more sensitive than PCR. Finally, LAMP-LFD gave no false positives across 350 chicken anal swabs. Therefore, our novel LAMP-LFD assay was highly sensitive, specific, convenient, and fast, making it a valuable tool for the early diagnosis and monitoring of Salmonella infection in chickens.

• The Journal of Korean Institute of Next Generation Computing
• /
• v.14 no.5
• /
• pp.60-70
• /
• 2018

A software defined storage (SDS) is being deployed in cloud environment to allow multiple users to virtualize physical servers, but a solution for optimizing space efficiency with limited physical resources is needed. In the conventional data deduplication system, it is difficult to deduplicate redundant data uploaded to distributed storages. In this paper, we propose a distributed deduplication method using similarity-based clustering and multi-layer bloom filter. Rabin hash is applied to determine the degree of similarity between virtual machine servers and cluster similar virtual machines. Therefore, it improves the performance compared to deduplication efficiency for individual storage nodes. In addition, a multi-layer bloom filter incorporated into the deduplication process to shorten processing time by reducing the number of the false positives. Experimental results show that the proposed method improves the deduplication ratio by 9% compared to deduplication method using IP address based clusters without any difference in processing time.

• Journal of Broadcast Engineering
• /
• v.26 no.1
• /
• pp.14-25
• /
• 2021

Change detection, which finds changes in remote sensing images of the same location captured at different times, is very important because it is used in various applications. However, registration errors, building displacement errors, and shadow errors cause false positives. To solve these problems, we propose a novle deep convolutional network called CADNet (Change Attention Dense Siamese Network). CADNet uses FPN (Feature Pyramid Network) to detect multi-scale changes, applies a Change Attention Module that attends to the changes, and uses DenseNet as a feature extractor to use feature maps that contain both low-level and high-level features for change detection. CADNet performance measured from the Precision, Recall, F1 side is 98.44%, 98.47%, 98.46% for WHU datasets and 90.72%, 91.89%, 91.30% for LEVIR-CD datasets. The results of this experiment show that CADNet can offer better performance than any other traditional change detection method.

• Biomedical Science Letters
• /
• v.27 no.1
• /
• pp.35-43
• /
• 2021

Human Sapovirus (HuSaV) is one of the major causes of acute gastroenteritis in humans, and it is used as a molecular diagnostic technique based on polymerase chain reaction (PCR) from humans, food, shellfish, and aquatic environments. In this study, the HuSaV diagnosis technique was used in an aquatic environment where a number of PCR inhibitors are included and pathogens, such as viruses, are estimated to exist at low concentration levels. HuSaV-specific primers are improved to detect 38 strains registered in the National Center for Biotechnology Information (NCBI). The established optimal condition and the composition, including the RT-nested PCR primers and SL® Non-specific reaction inhibitor, were found to have 100 times higher sensitivity based on HuSaV plasmid than the previously reported methods (100 ag based on HuSaV plasmid 1 ng/μL). Through an artificial infection test, the developed method was able to detect at least 1 fg/μL of HuSaV plasmid contaminated with total nucleic acid extracted from groundwater. In addition, RT-nested PCR primer sets for HuSaV detection can react, and a positive control is developed to verify false positives. This study is expected to be used as a HuSaV monitoring method in the future and applied to the safety response to HuSaV from water environments.

• Journal of the Korea Institute of Information and Communication Engineering
• /
• v.26 no.6
• /
• pp.907-913
• /
• 2022

Approximately 70% of smartphone users around the world use Android operating system-based smartphones, and malicious apps targeting these Android platforms are constantly increasing. Google has provided "Google Play Protect" to respond to the increasing number of Android targeted malware, preventing malicious apps from being installed on smartphones, but many malicious apps are still normal. It threatens the smartphones of ordinary users registered in the Google Play store by disguising themselves as apps. However, most people rely on antivirus programs to detect malicious apps because the average user needs a great deal of expertise to check for malicious apps. Therefore, in this paper, we propose a method to classify unnecessary malicious permissions of apps by using only the categories and permissions that can be easily confirmed by the app, and to easily detect malicious apps through the classified permissions. The proposed method is compared and analyzed from the viewpoint of undiscovered rate and false positives with the "commercial malicious application detection program", and the performance level is presented.

• KSII Transactions on Internet and Information Systems (TIIS)
• /
• v.16 no.7
• /
• pp.2131-2153
• /
• 2022

To achieve accurate detection of tuberculosis (TB) areas in chest radiographs, we design a chest X-ray TB area detection algorithm. The algorithm consists of two stages: the chest X-ray TB classification network (CXTCNet) and the chest X-ray TB area detection network (CXTDNet). CXTCNet is used to judge the presence or absence of TB areas in chest X-ray images, thereby excluding the influence of other lung diseases on the detection of TB areas. It can reduce false positives in the detection network and improve the accuracy of detection results. In CXTCNet, we propose a channel attention mechanism (CAM) module and combine it with DenseNet. This module enables the network to learn more spatial and channel features information about chest X-ray images, thereby improving network performance. CXTDNet is a design based on a sparse object detection algorithm (Sparse R-CNN). A group of fixed learnable proposal boxes and learnable proposal features are using for classification and location. The predictions of the algorithm are output directly without non-maximal suppression post-processing. Furthermore, we use CLAHE to reduce image noise and improve image quality for data preprocessing. Experiments on dataset TBX11K show that the accuracy of the proposed CXTCNet is up to 99.10%, which is better than most current TB classification algorithms. Finally, our proposed chest X-ray TB detection algorithm could achieve AP of 45.35% and AP50 of 74.20%. We also establish a chest X-ray TB dataset with 304 sheets. And experiments on this dataset showed that the accuracy of the diagnosis was comparable to that of radiologists. We hope that our proposed algorithm and established dataset will advance the field of TB detection.

• Journal of the Korea Institute of Information Security & Cryptology
• /
• v.31 no.6
• /
• pp.1193-1204
• /
• 2021

In information security, AI technology is used to detect unknown malware. Although AI technology guarantees high accuracy, it inevitably entails false positives, so we are considering introducing XAI to interpret the results predicted by AI. However, XAI evaluation studies that evaluate or verify the interpretation only provide simple interpretation results are lacking. XAI evaluation is essential to ensure safety which technique is more accurate. In this paper, we interpret AI results as features that have significantly contributed to AI prediction in the field of malware, and present an evaluation method for the interpretation of AI results. Interpretation of results is performed using two XAI techniques on a tree-based AI model with an accuracy of about 94%, and interpretation of AI results is evaluated by analyzing descriptive accuracy and sparsity. As a result of the experiment, it was confirmed that the AI result interpretation was properly calculated. In the future, it is expected that the adoption and utilization of XAI will gradually increase due to XAI evaluation, and the reliability and transparency of AI will be greatly improved.