• Journal of the Korea Society of Computer and Information
• /
• v.27 no.12
• /
• pp.59-68
• /
• 2022

This paper presents a method for 1:1 verification by comparing the similarity between the given real product image and the drawing image. The proposed method combines two existing CNN-based deep learning models to construct a Siamese Network. After extracting the feature vector of the image through the FC (Fully Connected) Layer of each network and comparing the similarity, if the real product image and the drawing image (front view, left and right side view, top view, etc) are the same product, the similarity is set to 1 for learning and, if it is a different product, the similarity is set to 0. The test (inference) model is a deep learning model that queries the real product image and the drawing image in pairs to determine whether the pair is the same product or not. In the proposed model, through a comparison of the similarity between the real product image and the drawing image, if the similarity is greater than or equal to a threshold value (Threshold: 0.5), it is determined that the product is the same, and if it is less than or equal to, it is determined that the product is a different product. The proposed model showed an accuracy of about 71.8% for a query to a product (positive: positive) with the same drawing as the real product, and an accuracy of about 83.1% for a query to a different product (positive: negative). In the future, we plan to conduct a study to improve the matching accuracy between the real product image and the drawing image by combining the parameter optimization study with the proposed model and adding processes such as data purification.

• Korean Journal of Environmental Biology
• /
• v.39 no.4
• /
• pp.581-589
• /
• 2021

Using visible-near infrared(Vis-NIR) spectra combined with machine learning methods, the feasibility of quick and non-destructive classification of Convolvulaceae species was studied. The main aim of this study is to classify six Convolvulaceae species in the field in different geographical regions of South Korea using a handheld spectrometer. Spectra were taken at 1.5 nm intervals from the adaxial side of the leaves in the Vis-NIR spectral region between 400 and 1,075 nm. The obtained spectra were preprocessed with three different preprocessing methods to find the best preprocessing approach with the highest classification accuracy. Preprocessed spectra of the six Convolvulaceae sp. were provided as input for the machine learning analysis. After cross-validation, the classification accuracy of various combinations of preprocessing and modeling ranged between 43.4% and 98.6%. The combination of Savitzky-Golay and Support vector machine methods showed the highest classification accuracy of 98.6% for the discrimination of Convolvulaceae sp. The growth stage of the plants, different measuring locations, and the scanning position of leaves on the plant were some of the crucial factors that affected the outcomes in this investigation. We conclude that Vis-NIR spectroscopy, coupled with suitable preprocessing and machine learning approaches, can be used in the field to effectively discriminate Convolvulaceae sp. for effective weed monitoring and management.

• Proceedings of the Korean Information Science Society Conference
• /
• 2011.06b
• /
• pp.197-199
• /
• 2011

테스팅은 테스트 요구사항을 기반으로 수행하기 때문에 테스트 요구사항의 품질은 테스팅 전체 과정의 품질과 직결된다고 할 수 있다. 그러나 테스트 요구사항에서 고려해야 할 점들은 시스템의 도메인과 목적에 따라 다르기 때문에 양질의 테스트 요구사항이라고 판별할 수 있는 기준을 정하는 것은 어려운 일이다. 본 연구에서는 테스트 케이스와 고장 수목의 최소 절단집합을 각각 정형모델로 변환하여 모델체킹을 함으로써, 테스트 요구사항의 안전성을 측정하는 방법을 제시한다. 테스트 요구사항이 반영된 테스트 케이스는 모델체킹의 대상이 되는 정형모델로 변환하였으며, 고장수목의 최소 절단집합은 CTL 검증 속성으로 변환하여 테스트 케이스에서 생성된 정형모델이 안전성을 만족하는지 만족하는지 모델체킹을 적용하여 확인하였다.

• Journal of Navigation and Port Research
• /
• v.31 no.8
• /
• pp.725-731
• /
• 2007

This paper aims to analyze an action about a port choice of shipper between two ports. For this propose, this paper analyzed a port choice action for Kwangyang and Busan port using a fuzzy logic and neural network. Also, this paper compared classification performance of fuzzy-neural network to Logit model, and analyzed a port choice action into change Para-meta such as freight volumes and service standard.

• Proceedings of the Korea Information Processing Society Conference
• /
• 2019.10a
• /
• pp.942-945
• /
• 2019

현재 딥러닝은 컴퓨터 분야에서 이미지 처리 방법으로 활용도가 높아지면서 딥러닝 모델 개발 연구가 활발히 진행되고 있다. 딥러닝 모델 중에서 이미지 생성모델은 대표적으로 GAN(Generative Adversarial Network, 생성적 적대 신경망) 모델을 활용하고 있다. GAN은 생성기 네트워크와 판별기 네트워크를 이용하여 진짜 같은 이미지를 생성한다. 생성된 이미지는 실제 이미지와의 오차를 최소화해야 하며 이때 사용하는 함수를 손실함수라고 한다. GAN에서 손실함수는 이미지를 생성하는 학습이 불안정하여 이미지 품질이 떨어지는 문제가 있다. 개선된 GAN 관련 연구가 진행되고 있지만 완전한 문제 해결에는 부족하다. 본 논문은 7개의 GAN 모델에서 사용하는 손실함수를 분류하고 특징을 분석한다.

• Proceedings of the Korea Information Processing Society Conference
• /
• 2022.05a
• /
• pp.524-527
• /
• 2022

본 연구에서는 머신러닝(Machine Learning, ML)과 딥러닝(Deep Learning, DL) 모델을 앙상블(Ensemble)하여 어떠한 주가 예측 방법이 우수한지에 대한 연구를 하고자 한다. 연구에 사용된 모델은 하이퍼파라미터(Hyperparameter) 조정을 통하여 최적의 결과를 출력한다. 앙상블 방법은 머신러닝과 딥러닝 모델의 앙상블, 머신러닝 모델의 앙상블, 딥러닝 모델의 앙상블이다. 세 가지 방법으로 얻은 결과를 평균 제곱근 오차(Root Mean Squared Error, RMSE)로 비교 분석하여 최적의 방법을 찾고자 한다. 제안한 방법은 주가 예측 연구의 시간과 비용을 절약하고, 최적 성능 모델 판별에 도움이 될 수 있다고 사료된다.

• Proceedings of the Korea Society for Simulation Conference
• /
• 2000.11a
• /
• pp.125-130
• /
• 2000

침입 탐지 시스템이 침입 행위를 탐지하기 위해서 시간에 대한 처리를 고려하지 않고는 침입을 탐지할 수 없는 경우(예 Denial of Service)가 존재한다. 즉 사건의 발생 시점에 대한 처리없이는 침입 탐지가 불가능하다. 본 논문에서는 시뮬레이션 모델을 통하여 시간에 관한 처리를 체계적으로 구성하고, 여러 가지 상황을 조성하여 반복적으로 실행함으로써 침입 탐지 시스템의 핵심 요소인 침입 판별을 효과적으로 수행할 수 있도록 하였다.

• Journal of the Society of Disaster Information
• /
• v.20 no.2
• /
• pp.302-314
• /
• 2024

Purpose: This study proposes a fall detection model based on a top-down deep learning pose estimation model to automatically determine falls of multiple workers in an underground utility tunnel, and evaluates the performance of the proposed model. Method: A model is presented that combines fall discrimination rules with the results inferred from YOLOv8-pose, one of the top-down pose estimation models, and metrics of the model are evaluated for images of standing and falling two or fewer workers in the tunnel. The same process is also conducted for a bottom-up type of pose estimation model (OpenPose). In addition, due to dependency of the falling interference of the models on worker detection by YOLOv8-pose and OpenPose, metrics of the models for fall was not only investigated, but also for person. Result: For worker detection, both YOLOv8-pose and OpenPose models have F1-score of 0.88 and 0.71, respectively. However, for fall detection, the metrics were deteriorated to 0.71 and 0.23. The results of the OpenPose based model were due to partially detected worker body, and detected workers but fail to part them correctly. Conclusion: Use of top-down type of pose estimation models would be more effective way to detect fall of workers in the underground utility tunnel, with respect to joint recognition and partition between workers.

• Journal of Internet of Things and Convergence
• /
• v.6 no.2
• /
• pp.1-10
• /
• 2020

The mobile application based on the Android platform is simple to decompile, making it possible to create malicious applications similar to normal ones, and can easily distribute the created malicious apps through the Android third party app store. In this case, the Android malicious application in the smartphone causes several problems such as leakage of personal information in the device, transmission of premium SMS, and leakage of location information and call records. Therefore, it is necessary to select a optimal model that provides the best performance among the machine learning techniques that have published recently, and provide a technique to automatically identify malicious Android apps. Therefore, in this paper, after adopting the feature engineering to Android apps on official test set, a total of four performance evaluation experiments were conducted to select the machine learning model that provides the optimal performance for Android malicious app detection.

• Journal of the Korea Institute of Information Security & Cryptology
• /
• v.26 no.5
• /
• pp.1259-1267
• /
• 2016

With the growing use of fingerprint authentication systems in recent years, the fake fingerprint detection is becoming more and more important. This paper mainly proposes a method for fake fingerprint detection based on CNN, it will visualize the distinctive part of detected fingerprint which provides a deeper insight in CNN model. After the preprocessing part using fingerprint segmentation, the pretrained CNN model is used for detecting the liveness detection. Not only a liveness detection but also feature analysis about the live fingerprint and fake fingerprint are provided after classifying which materials are used for making the fake fingerprint. Our system is evaluated on three databases in LivDet2013, which compromise almost 6500 live fingerprint images and 6000 fake fingerprint images in total. The proposed method achieves 3.1% ACE value about the liveness detection and achieves 79.58% accuracy on LiveDet2013.