• Journal of the Society of Disaster Information
• /
• 제18권2호
• /
• pp.364-373
• /
• 2022

Purpose: The purpose of this paper is to develope smoke detection using AI model for detecting the initial fire in underground utility tunnels using CCTV Method: To improve detection performance of smoke which is high irregular, a deep learning model for fire detection was trained to optimize smoke detection. Also, several approaches such as dataset cleansing and gradient exploding release were applied to enhance model, and compared with results of those. Result: Results show the proposed approaches can improve the model performance, and the final model has good prediction capability according to several indexes such as mAP. However, the final model has low false negative but high false positive capacities. Conclusion: The present model can apply to smoke detection in underground utility tunnel, fixing the defect by linking between the model and the utility tunnel control system.

• Korean Journal of Remote Sensing
• /
• 제38권5_1호
• /
• pp.559-570
• /
• 2022

When a satellite image has low spatial resolution, it is difficult to detect small objects. In this research, we aim to check the effect of super resolution on object detection. Super resolution is a software method that increases the resolution of an image. Unpaired super resolution network is used to improve Sentinel-2's spatial resolution from 10 m to 3.2 m. Faster-RCNN, RetinaNet, FCOS, and S²ANet were used to detect vessels in the Sentinel-2 images. We experimented the change in vessel detection performance when super resolution is applied. As a result, the Average Precision (AP) improved by at least 12.3% and up to 33.3% in the ship detection models trained with the super-resolution image. False positive and false negative cases also decreased. This implies that super resolution can be an important pre-processing step in object detection, and it is expected to greatly contribute to improving the accuracy of other image-based deep learning technologies along with object detection.

• KSII Transactions on Internet and Information Systems (TIIS)
• /
• 제16권12호
• /
• pp.4008-4023
• /
• 2022

Since machine learning was introduced into cross-site scripting (XSS) attack detection, many researchers have conducted related studies and achieved significant results, such as saving time and labor costs by not maintaining a rule database, which is required by traditional XSS attack detection methods. However, this topic came across some problems, such as poor generalization ability, significant false negative rate (FNR) and false positive rate (FPR). Moreover, the automatic clustering property of graph convolutional networks (GCN) has attracted the attention of researchers. In the field of natural language process (NLP), the results of graph embedding based on GCN are automatically clustered in space without any training, which means that text data can be classified just by the embedding process based on GCN. Previously, other methods required training with the help of labeled data after embedding to complete data classification. With the help of the GCN auto-clustering feature and labeled data, this research proposes an approach to detect XSS attacks (called GCNXSS) to mine the dependencies between the units that constitute an XSS payload. First, GCNXSS transforms a URL into a word homogeneous graph based on word co-occurrence relationships. Then, GCNXSS inputs the graph into the GCN model for graph embedding and gets the classification results. Experimental results show that GCNXSS achieved successful results with accuracy, precision, recall, F1-score, FNR, FPR, and predicted time scores of 99.97%, 99.75%, 99.97%, 99.86%, 0.03%, 0.03%, and 0.0461ms. Compared with existing methods, GCNXSS has a lower FNR and FPR with stronger generalization ability.

• Journal of Gastric Cancer
• /
• 제22권4호
• /
• pp.395-407
• /
• 2022

Purpose: This study aimed to investigate the relationship between clinical and laboratory parameters and complication status to predict which patients can be safely discharged from the hospital on the third postoperative day (POD). Materials and Methods: Data from a prospectively maintained database of 2,110 consecutive patients with gastric adenocarcinoma who underwent curative surgery were reviewed. The third POD vital signs, laboratory data, and details of the course after surgery were collected. Patients with grade II or higher complications after the third POD were considered unsuitable for early discharge. The performance metrics were calculated for all algorithm parameters. The proposed algorithm was tested using a validation dataset of consecutive patients from the same center. Results: Of 1,438 patients in the study cohort, 142 (9.9%) were considered unsuitable for early discharge. C-reactive protein level, body temperature, pulse rate, and neutrophil count had good performance metrics and were determined to be independent prognostic factors. An algorithm consisting of these 4 parameters had a negative predictive value (NPV) of 95.9% (95% confidence interval [CI], 94.2-97.3), sensitivity of 80.3% (95% CI, 72.8-86.5), and specificity of 51.1% (95% CI, 48.3-53.8). Only 28 (1.9%) patients in the study cohort were classified as false negatives. In the validation dataset, the NPV was 93.7%, sensitivity was 66%, and 3.3% (17/512) of patients were classified as false negatives. Conclusions: Simple clinical and laboratory parameters obtained on the third POD can be used when making decisions regarding the safe early discharge of patients who underwent gastrectomy.

• Journal of Industrial Convergence
• /
• 제20권11호
• /
• pp.191-196
• /
• 2022

In the fintech field, financial transactions with smart phones are actively conducted. User authentication technology is essential for safe financial transactions. PIN authentication through the existing security keypads is convenient to input but has weaknesses in security and others. The biometric authentication technique is secure, but there is a possibility of false positive and false negative authentication. To compensate for this, two-factor authentication is used. In this paper, we propose the 1.5-factor authentication that can increase convenience and security through PIN input with biometric authentication. It provides the stability of fingerprint authentication and convenience of two or three PIN inputs, and this makes safe financial transaction possible. Since biometric authentication is performed at the same time when entering PIN, while security is required by applying fingerprint authentication to the area touched while entering PIN. The User authentication is performed while ensuring convenience to input through additional PIN input in situations where high safety is required, and Safe financial transactions are possible.

• Asian Pacific Journal of Cancer Prevention
• /
• 제16권8호
• /
• pp.3229-3232
• /
• 2015

Background: Triple-negative breast cancer (TNBC) is known to be associated with aggressive biologic features and a poor clinical outcome. Therefore, early detection of TNBC without missed diagnosis is a requirement to improve prognosis. Preoperative ultrasound features of TNBC may potentially assist in early diagnosis as characteristics of disease. Purpose: To retrospectively evaluate the sonographic features of TNBC compared to ER (+) cancers which include HER(-) and HER2 (+), and HER2 (+) cancers which are ER (-). Materials and Methods: From June 2012 through June 2014, sonographic features of 321 surgically confirmed ER (+) cancers (n=214), HER2 (+) cancers (n=66), and TNBC (n=41) were retrospectively reviewed by two ultrasound specialists in consensus. The preoperative ultrasound and clinicopathological features were compared between the three subtypes. In addition, all cases were analyzed using morphologic criteria of the ACR BI-RADS lexicon. Results: Ultrasonographically, TNBC presented as microlobulated nodules without microcalcification (p=0.034). A lower incidence of ductal carcinoma in situ (p<0.001), invasive tumor size that is>2 cm (p=0.011) and BI-RADS category 4 (p<0.001) were significantly associated with TNBC. With regard to morphologic features of 41 TNBC cases, ultrasonographically were most likely to be masses with irregular (70.7%) microlobulated shape (48.8%), be circumscribed (17.1%) or have indistinct margins (17.1%) and parallel orientation (68.9%). Especially TNBC microlobulated mass margins were more more frequent than with ER (+) (2.0%) and HER2 (+) (4.8%) cancers. Conclusions: TNBC have specific characteristic in sonograms. Ultrasonography may be useful to avoid missed diagnosis and false-negative cases of TNBC.

• Parasites, Hosts and Diseases
• /
• 제53권2호
• /
• pp.147-154
• /
• 2015

Various constituents in clinical specimens, particularly feces, can inhibit the PCR assay and lead to false-negative results. To ensure that negative results of a diagnostic PCR assay are true, it should be properly monitored by an inhibition control. In this study, a cloning vector harboring a modified target DNA sequence (${\approx}375bp$) was constructed to be used as a competitive internal amplification control (IAC) for a conventional PCR assay that detects ${\approx}550bp$ of the Cryptosporidium oocyst wall protein (COWP) gene sequence in human feces. Modification of the native PCR target was carried out using a new approach comprising inverse PCR and restriction digestion techniques. IAC was included in the assay, with the estimated optimum concentration of 1 fg per reaction, as duplex PCR. When applied on fecal samples spiked with variable oocysts counts, ${\approx}2$ oocysts were theoretically enough for detection. When applied on 25 Cryptosporidium-positive fecal samples of various infection intensities, both targets were clearly detected with minimal competition noticed in 2-3 samples. Importantly, both the analytical and the diagnostic sensitivities of the PCR assay were not altered with integration of IAC into the reactions. When tried on 180 randomly collected fecal samples, 159 were Cryptosporidium-negatives. Although the native target DNA was absent, the IAC amplicon was obviously detected on gel of all the Cryptosporidium-negative samples. These results imply that running of the diagnostic PCR, inspired with the previously developed DNA extraction protocol and the constructed IAC, represents a useful tool for Cryptosporidium detection in human feces.

• Journal of Intelligence and Information Systems
• /
• 제25권1호
• /
• pp.127-137
• /
• 2019

Recidivism prediction has been a subject of constant research by experts since the early 1970s. But it has become more important as committed crimes by recidivist steadily increase. Especially, in the 1990s, after the US and Canada adopted the 'Recidivism Risk Assessment Report' as a decisive criterion during trial and parole screening, research on recidivism prediction became more active. And in the same period, empirical studies on 'Recidivism Factors' were started even at Korea. Even though most recidivism prediction studies have so far focused on factors of recidivism or the accuracy of recidivism prediction, it is important to minimize the prediction misclassification cost, because recidivism prediction has an asymmetric error cost structure. In general, the cost of misrecognizing people who do not cause recidivism to cause recidivism is lower than the cost of incorrectly classifying people who would cause recidivism. Because the former increases only the additional monitoring costs, while the latter increases the amount of social, and economic costs. Therefore, in this paper, we propose an XGBoost(eXtream Gradient Boosting; XGB) based recidivism prediction model considering asymmetric error cost. In the first step of the model, XGB, being recognized as high performance ensemble method in the field of data mining, was applied. And the results of XGB were compared with various prediction models such as LOGIT(logistic regression analysis), DT(decision trees), ANN(artificial neural networks), and SVM(support vector machines). In the next step, the threshold is optimized to minimize the total misclassification cost, which is the weighted average of FNE(False Negative Error) and FPE(False Positive Error). To verify the usefulness of the model, the model was applied to a real recidivism prediction dataset. As a result, it was confirmed that the XGB model not only showed better prediction accuracy than other prediction models but also reduced the cost of misclassification most effectively.

• Journal of Korean Society of Forest Science
• /
• 제105권4호
• /
• pp.422-428
• /
• 2016

Ginkgo (Ginkgo biloba L.) is one of the most appropriate roadside trees because of a good transplantation nature and ability to grow well in urban environment. Ginkgo is a dioecious species. Sex discrimination of ginkgo is possible through comparing morphological characters of reproductive organs. However, it needs more than about twenty years for reproductive organs to appear after sexual maturity. Until now, ginkgo trees for roadside plantation have been planted without discriminating the sex because ginkgo trees have been usually planted before sexual maturity. Ginkgo nuts from the female ginkgo trees planted along the roadside emit a foul odor, and make much pollution on the streets. Thus in this study a novel SCAR marker (SCAR-GBM) for the early sex discrimination was developed. Primers were developed on the basis of the sequence of male-specific RAPD variants reported previously. False-negative problem of SCAR marker, probably caused by dominant nature, was resolved by using multiplex PCR using primers of both the SCAR-GBM and a universal primer set of atp1 region in mitochondria DNA, which resulted in improved discrimination efficiency. The results showed that DNA bands of 1,039 bp were commonly amplified by the atp1 primer set in male and female trees, and SCAR-GBM markers of 675 bp were specifically amplified only in male trees. Reproducible and specific discrimination of the multiplex PCR was finally confirmed by applying multiple male and female individuals.

• Journal of Food Hygiene and Safety
• /
• 제33권1호
• /
• pp.65-70
• /
• 2018

Monitoring of Listeria monocytogenes, the causative agent of listeriosis, in food is inportant for public health. The Korean Food Standards Codex has adopted a 'zero-tolerance' policy for L. monocytogenes. The standard detection method of L. monocytogenes is based on enrichment. Thus, proper enrichment methods need to be instituted to ensure quality control of the detection procedures. In this study, the growth of L. monocytogenes and Listeria innocua as a mixed culture in Listeria enrichment broth (LEB) was monitored during artificial contamination of enrichment culture. We confirmed competitive growth or interspecies inhibitory activity of L. monocytogenes and L. innocua. Interspecies growth differences and the inhibitory activity of different inoculation and mixtures L. innocua against L. monocytogenes were examined. The concentration of L. monocytogenes must be 2.0 log CFU/mL or more than L. innocua to grow better than L. innocua. It is known that Listeria spp. and L. monocytogenes show growth difference during LEB, resulting in the risk of false-negative results. The inhibition of L. monocytogenes by L. innocua was always observed when present at lower concentrations. However, it was confirmed that L. innocua suppressed when L. monocytogenes was present at a higher concentration. Therefore if a mixture of Listeria spp. is present, detecting L. monocytogenes is difficult. Thus, a new enrichment broth to improve the detection rate of L. monocytogenes is needed.