• Convergence Security Journal
• /
• v.23 no.5
• /
• pp.135-142
• /
• 2023

In addressing the prominent issues of climate change, resource scarcity, and environmental pollution associated with household waste, extensive research has been conducted on intelligent waste classification methods. These efforts range from traditional classification algorithms to machine learning and neural networks. However, challenges persist in effectively classifying waste in diverse environments and conditions due to insufficient datasets, increased complexity in neural network architectures, and performance limitations for real-world applications. Therefore, this paper proposes i-YOLOX as a solution for rapid classification and improved accuracy. The proposed model is evaluated based on network parameters, detection speed, and accuracy. To achieve this, a dataset comprising 10,000 samples of household waste, spanning 17 waste categories, is created. The i-YOLOX architecture is constructed by introducing the Involution channel convolution operator and the Convolution Branch Attention Module (CBAM) into the YOLOX structure. A comparative analysis is conducted with the performance of the existing YOLO architecture. Experimental results demonstrate that i-YOLOX enhances the detection speed and accuracy of waste objects in complex scenes compared to conventional neural networks. This confirms the effectiveness of the proposed i-YOLOX architecture in the detection and classification of multiple household waste objects.

• Smart Media Journal
• /
• v.12 no.11
• /
• pp.1-9
• /
• 2023

Stress is a significant issue in modern society, often triggered by external or internal factors that are difficult to manage. When high stress persists over a long term, it can develop into a chronic condition, negatively impacting health and overall well-being. However, it is challenging for individuals experiencing chronic stress to recognize their condition, making early detection and management crucial. Using biosignals measured from wearable devices to detect stress could lead to more effective management. However, there are two main problems with using biosignals: first, manually extracting features from these signals can introduce bias, and second, the performance of classification models can vary greatly depending on the subject of the experiment. This paper proposes a model that reduces bias using convo utional autoencoders, which can represent the key features of data, and enhances generalizability by employing soft voting, a method of ensemble learning, to minimize performance variability. To verify the generalization performance of the model, we evaluate it using LOSO cross-validation method. The model proposed in this paper has demonstrated superior accuracy compared to previous studies using the WESAD dataset.

• KIPS Transactions on Computer and Communication Systems
• /
• v.5 no.10
• /
• pp.293-302
• /
• 2016

Since the electricity is produced and consumed simultaneously, predicting the electric load and securing affordable electric power are necessary for reliable electric power supply. In particular, a university campus is one of the highest power consuming institutions and tends to have a wide variation of electric load depending on time and environment. For these reasons, an accurate electric load forecasting method that can predict power consumption in real-time is required for efficient power supply and management. Even though various influencing factors of power consumption have been discovered for the educational institutions by analyzing power consumption patterns and usage cases, further studies are required for the quantitative prediction of electric load. In this paper, we build an electric load forecasting model by implementing and evaluating various machine learning algorithms. To do that, we consider three building clusters in a campus and collect their power consumption every 15 minutes for more than one year. In the preprocessing, features are represented by considering periodic characteristic of the data and principal component analysis is performed for the features. In order to train the electric load forecasting model, we employ both artificial neural network and support vector machine. We evaluate the prediction performance of each forecasting model by 5-fold cross-validation and compare the prediction result to real electric load.

• The Journal of The Korea Institute of Intelligent Transport Systems
• /
• v.21 no.1
• /
• pp.105-122
• /
• 2022

According to the statistics about the fatal crashes that have occurred on the expressways for the last 5 years, those who died on the shoulders of the road has been as 3 times high as the others who died on the expressways. It suggests that the crashes on the shoulders of the road should be fatal, and that it would be important to prevent the traffic crashes by cracking down on the vehicles intruding the shoulders of the road. Therefore, this study proposed a method to detect a vehicle that violates the shoulder lane by using the Faster R-CNN. The vehicle was detected based on the Faster R-CNN, and an additional reading module was configured to determine whether there was a shoulder violation. For experiments and evaluations, GTAV, a simulation game that can reproduce situations similar to the real world, was used. 1,800 images of training data and 800 evaluation data were processed and generated, and the performance according to the change of the threshold value was measured in ZFNet and VGG16. As a result, the detection rate of ZFNet was 99.2% based on Threshold 0.8 and VGG16 93.9% based on Threshold 0.7, and the average detection speed for each model was 0.0468 seconds for ZFNet and 0.16 seconds for VGG16, so the detection rate of ZFNet was about 7% higher. The speed was also confirmed to be about 3.4 times faster. These results show that even in a relatively uncomplicated network, it is possible to detect a vehicle that violates the shoulder lane at a high speed without pre-processing the input image. It suggests that this algorithm can be used to detect violations of designated lanes if sufficient training datasets based on actual video data are obtained.

• Journal of the Korean Society of Surveying, Geodesy, Photogrammetry and Cartography
• /
• v.37 no.6
• /
• pp.435-444
• /
• 2019

Recently, DL (Deep Learning) has been rapidly applied in various fields. In particular, classification and object recognition from images are major tasks in computer vision. Most of the DL utilizing imagery is primarily based on the CNN (Convolutional Neural Network) and improving performance of the DL model is main issue. While most CNNs are involve with images for training data, this paper aims to classify and recognize objects using DSM (Digital Surface Model), and slope and aspect information derived from the DSM instead of images. The DSM data sets used in the experiment were established by DGPF (German Society for Photogrammetry, Remote Sensing and Geoinformatics) and provided by ISPRS (International Society for Photogrammetry and Remote Sensing). The CNN-based SegNet model, that is evaluated as having excellent efficiency and performance, was used to train the data sets. In addition, this paper proposed a scheme for training data generation efficiently from the limited number of data. The results demonstrated DSM and derived data could be feasible for semantic classification with desirable accuracy using DL.

• KIPS Transactions on Computer and Communication Systems
• /
• v.11 no.7
• /
• pp.233-240
• /
• 2022

Most of the recent AI researches has focused on developing AI models. However, recently, artificial intelligence research has gradually changed from model-centric to data-centric, and the importance of learning data is getting a lot of attention based on this trend. However, it takes a lot of time and effort because the preparation of learning data takes up a significant part of the entire process, and the generation of labeling data also differs depending on the purpose of development. Therefore, it is need to develop a tool with various labeling functions to solve the existing unmetneeds. In this paper, we describe a labeling system for creating precise and fast labeling data of medical images. To implement this, a semi-automatic method using Back Projection, Grabcut techniques and an automatic method predicted through a machine learning model were implemented. We not only showed the advantage of running time for the generation of labeling data of the proposed system, but also showed superiority through comparative evaluation of accuracy. In addition, by analyzing the image data set of about 1,000 patients, meaningful diagnostic indexes were presented for men and women in the diagnosis of sarcopenia.

• The Journal of the Convergence on Culture Technology
• /
• v.9 no.6
• /
• pp.301-307
• /
• 2023

Deep learning requires building a deep neural network, collecting a large amount of training data, and then training the built neural network for a long time. If training does not proceed properly or overfitting occurs, training will fail. When using deep learning tools that have been developed so far, it takes a lot of time to collect training data and learn. However, due to the rapid advent of the mobile environment and the increase in sensor data, the demand for real-time deep learning technology that can dramatically reduce the time required for neural network learning is rapidly increasing. In this study, a real-time deep learning system was implemented using an Arduino system equipped with a fine dust sensor. In the implemented system, fine dust data is measured every 30 seconds, and when up to 120 are accumulated, learning is performed using the previously accumulated data and the newly accumulated data as a dataset. The neural network for learning was composed of one input layer, one hidden layer, and one output. To evaluate the performance of the implemented system, learning time and root mean square error (RMSE) were measured. As a result of the experiment, the average learning error was 0.04053796, and the average learning time of one epoch was about 3,447 seconds.

• Proceedings of the Korean Information Science Society Conference
• /
• 2006.06d
• /
• pp.235-237
• /
• 2006

최근 분산 해쉬 테이블(Distributed Hash Table)을 이용한 구조적 P2P(peer-to-peer) 시스템에서 로드 밸런싱에 관한 연구가 활발하게 진행되고 있다. 일반적으로 P2P 시스템에서의 피어들은 빈번하게 참여하는 동시에 탈퇴하기도 하기 때문에 로드의 불균형이 심화되는 문제점이 있다. 본 논문에서는 구조적 P2P 시스템에서 피어의 참여와 탈퇴 시 발생하는 가상서버의 이동을 효과적으로 관리하는 기법인 PALB(Peer Activity-aware Load Balancing)를 제안한다. 제안하는 PALB는 피어의 참여, 탈퇴 시에 발생하는 로드 불균형을 해결하기 위하여 적절한 수의 가상서버를 가장 적합한 피어에게 분산시킨다. 또한 이전 연구에서 제안한 OLAB(Object Lifetime-aware Load Balancing)와 함께 사용하여 로드 밸런싱 성능을 향상시킬 수 있다. PALB의 성능 평가를 위하여 동적 P2P 시스템 환경을 구성하였고, 피어와 오브젝트의 랜덤 데이터 셋을 이용하여 시뮬레이션을 수행하였다. 시뮬레이션 결과 PALB가 기존의 시스템에 비하여 더 적은 비용으로 더 나은 로드 밸런싱을 수행함을 확인하였다. 또한 OLAB와 동시에 사용하였을 경우, 매우 뛰어난 성능을 보임을 확인하였다.

• Journal of KIISE
• /
• v.44 no.2
• /
• pp.179-185
• /
• 2017

Document summarization aims to generate a summary that is consistent and contains the highly related sentences in a document. In this study, we implemented for document summarization that extracts highly related sentences from a whole document by considering both similarities and entailment relations between sentences. Accordingly, we proposed a new algorithm, TextRank-NLI, which combines a Recurrent Neural Network based Natural Language Inference model and a Graph-based ranking algorithm used in single document extraction-based summarization task. In order to evaluate the performance of the new algorithm, we conducted experiments using the same datasets as used in TextRank algorithm. The results indicated that TextRank-NLI showed 2.3% improvement in performance, as compared to TextRank.

• Proceedings of the Korean Society of Broadcast Engineers Conference
• /
• 2011.11a
• /
• pp.154-157
• /
• 2011

인간의 동작 인식은 가상 현실 시스템 및 게임 등에 적용할 수 있는 컴퓨터 비전 분야의 요소 기술 중 하나로써, 최근까지 그 연구과 활발히 진행되고 있다. 본 논문에서는 빠르고 정확한 동작 인식을 위해, 실루엣과 모션 특징이 결합된 방법을 제안한다. 제안하는 방법은 전역 특징을 이용한 후보 동작 선정 및 지역 특징을 이용한 검증 2 단계로 구성된다. 전역 특징은 Motion History Image의 Hu 모멘트를 이용해 계산되며, 후보 동작의 선정은 이들의 통계치를 이용해 결정한다. 한정된 후보 동작들 중 입력 동작을 정확히 인식하기 위해, 공간 및 방향성 비닝 기법으로 추출된 광류와 실루엣 특징을 지역 특징으로 이용한다. 최종 인식 결과는 Hu 모멘트 통계치와의 유사도 및 지역 특징의 학습을 통해 생성된 Support Vector Machine의 결과를 고려하여 결정된다. 제안하는 방법의 성능을 평가하기 위해, 실세계에서 사용 빈도가 높으며 동작의 변화가 큰 13 개의 제스처를 선정하여 데이터 셋을 구성하였다. 실험 결과 제안하는 방법의 연산 시간은 50 ms, 인식 정확도는 95%임을 확인하였다.