• Annual Conference of KIPS
• /
• 2024.10a
• /
• pp.920-921
• /
• 2024

본 논문은 스마트 리테일 환경에서 객체 검지 기술의 적용에 관한 방법을 제안한다. 스마트 리테일은 자영업자나 소상공인 사업장의 전반적인 운영과 고객의 요구를 충족하기 위하여 첨단 기술을 활용하는 분야이다. 객체 검지 기술은 영상에서 사람, 물체 등 다양한 객체를 실시간으로 검지하여 객체 종류, 위치와 크기 정보 등을 제공하며, 이러한 정보는 재고관리, 고객 행동 분석, 무인 결제 시스템 등에서 유용하게 활용될 수 있다. 최근에는 딥러닝 기반의 객체 검지 기술이 우수한 성능을 보이고 다양한 분야에서 활용되고 있지만, 고성능의 객체 검지 기술을 운용하기 위해 고가의 하드웨어가 필요하다는 문제가 존재한다. 본 논문에서는 웹기반의 스마트 리테일 서비스 모델과 소규모사업장환경의 저가 하드웨어에서 객체 검지 기술을 적용하는 방법을 제안하고, 스마트 리테일에서 가장 중요한 객체인 사람 검지에 관한 성능을 분석한다.

• Journal of Korea Water Resources Association
• /
• v.57 no.7
• /
• pp.437-449
• /
• 2024

Rainfall-runoff prediction studies using deep learning while considering catchment attributes have been gaining attention. In this study, we selected two models: the Transformer model, which is suitable for large-scale data training through the self-attention mechanism, and the LSTM-based multi-state-vector sequence-to-sequence (LSTM-MSV-S2S) model with an encoder-decoder structure. These models were constructed to incorporate catchment attributes and predict the inflow of 10 multi-purpose dam watersheds in South Korea. The experimental design consisted of three training methods: Single-basin Training (ST), Pretraining (PT), and Pretraining-Finetuning (PT-FT). The input data for the models included 10 selected watershed attributes along with meteorological data. The inflow prediction performance was compared based on the training methods. The results showed that the Transformer model outperformed the LSTM-MSV-S2S model when using the PT and PT-FT methods, with the PT-FT method yielding the highest performance. The LSTM-MSV-S2S model showed better performance than the Transformer when using the ST method; however, it showed lower performance when using the PT and PT-FT methods. Additionally, the embedding layer activation vectors and raw catchment attributes were used to cluster watersheds and analyze whether the models learned the similarities between them. The Transformer model demonstrated improved performance among watersheds with similar activation vectors, proving that utilizing information from other pre-trained watersheds enhances the prediction performance. This study compared the suitable models and training methods for each multi-purpose dam and highlighted the necessity of constructing deep learning models using PT and PT-FT methods for domestic watersheds. Furthermore, the results confirmed that the Transformer model outperforms the LSTM-MSV-S2S model when applying PT and PT-FT methods.

• Journal of the Korea Convergence Society
• /
• v.9 no.4
• /
• pp.23-29
• /
• 2018

To reduce the accident rate caused by the delay of corrective action, which is common in the construction site, in order to shorten the time from correcting the existing system to the corrective action, I used a time similarity check to inform the inspectors of the problem in real time, modeling the system so that corrective action can be performed immediately on site, and studied a system that can actively cope with safety accidents. The research result shows that there is more than 90% opening effect and more than 60% safety accident reduction rate. I will continue to study more effective system combining voice recognition and deep learning based on this system.

• KIISE Transactions on Computing Practices
• /
• v.23 no.7
• /
• pp.440-445
• /
• 2017

Due to the advance of e-commerce systems, the number of people using online shopping and products has significantly increased. Therefore, the need for an accurate recommendation system is becoming increasingly more important. Recurrent neural network is a deep-learning algorithm that utilizes sequential information in training. In this paper, an evaluation is performed on the application of recurrent neural networks to recommendation systems. We evaluated three recurrent algorithms (RNN, LSTM and GRU) and three optimal algorithms(Adagrad, RMSProp and Adam) which are commonly used. In the experiments, we used the TensorFlow open source library produced by Google and e-commerce session data from RecSys Challenge 2015. The results using the optimal hyperparameters found in this study are compared with those of RecSys Challenge 2015 participants.

• The Journal of the Convergence on Culture Technology
• /
• v.9 no.3
• /
• pp.731-736
• /
• 2023

Recently, the medical field has been applying mandatory Electronic Medical Records (EMRs) and Electronic Health Records (EHRs) systems that computerize and manage medical records, and distributing them throughout the entire medical industry to utilize patients' past medical records for additional medical procedures. However, the conversations between medical professionals and patients that occur during general medical consultations and counseling sessions are not separately recorded or stored, so additional important patient information cannot be efficiently utilized. Therefore, we propose an electronic medical record system that uses speech recognition and natural language processing deep learning to store conversations between medical professionals and patients in text form, automatically extracts and summarizes important medical consultation information, and generates electronic medical records. The system acquires text information through the recognition process of medical professionals and patients' medical consultation content. The acquired text is then divided into multiple sentences, and the importance of multiple keywords included in the generated sentences is calculated. Based on the calculated importance, the system ranks multiple sentences and summarizes them to create the final electronic medical record data. The proposed system's performance is verified to be excellent through quantitative analysis.

• Proceedings of the Korean Institute of Information and Commucation Sciences Conference
• /
• 2018.05a
• /
• pp.464-465
• /
• 2018

Bass and Bluegill, which are representative ecosystem disturbance species, are reported to be the most important factor in the reduction of domestic native fish populations in Korea. Therefore, it is necessary to develop system and field application technology for the extermination of these foreign species. Recently, the CNN(Convolutional Neural Network), one of the deep learning systems for the recognition, classification, and learning, has shown excellent performance. However, CNN data used for object recognition and classification were mainly applied to recognition and classification of other objects with distinct characteristics. This study proposes a system that applies CNN to the classification of fish species with similar characteristics.

• Journal of Digital Convergence
• /
• v.18 no.11
• /
• pp.241-246
• /
• 2020

In this paper, we propose a neural network model to converge the marine electric propulsion system and deep learning algorithm to predict the DC/DC converter output current in the electric propulsion regeneration system and to predict the battery charge during regeneration. In order to experiment with the proposed neural network, the input voltage and current of the PCM were measured and the data set was secured on the prototype PCM board. In addition, in order to improve the learning results in the insufficient data set, the scale of the data set was increased through data fitting and its learning was executed further. After learning, the difference between the data prediction result of the neural network model and the actual measurement data was compared. The proposed neural network model effectively showed the prediction of battery charge according to changes in input voltage and current. In addition, by predicting the characteristic change of the analog circuit constituting the DC/DC converter through a neural network, it is determined that the characteristics of the analog circuit should be considered when designing the regeneration system.

• Journal of the Korea Convergence Society
• /
• v.11 no.5
• /
• pp.1-8
• /
• 2020

Machine translation refers to a system where a computer translates a source sentence into a target sentence. There are various subfields of machine translation. APE (Automatic Post Editing) is a subfield of machine translation that produces better translations by editing the output of machine translation systems. In other words, it means the process of correcting errors included in the translations generated by the machine translation system to make proofreading. Rather than changing the machine translation model, this is a research field to improve the translation quality by correcting the result sentence of the machine translation system. Since 2015, APE has been selected for the WMT Shaed Task. and the performance evaluation uses TER (Translation Error Rate). Due to this, various studies on the APE model have been published recently, and this paper deals with the latest research trends in the field of APE.

• Journal of Digital Convergence
• /
• v.16 no.12
• /
• pp.301-307
• /
• 2018

The most of the IR focus on the method for searching the document, so the keyword-based IR system is not able to reflect the feature information of the image. In order to overcome these limitations, we have developed a system that can search similar images based on the vector information of images, and it can search for similar images based on sketches. The proposed system uses the GAN to up sample the sketch to the image level, convert the image to the vector through the CNN, and then retrieve the similar image using the vector space model. The model was learned using fashion image and the image retrieval system was developed. As a result, the result is showed meaningful performance.

• Convergence Security Journal
• /
• v.21 no.1
• /
• pp.93-100
• /
• 2021

Damage caused by intelligent cyber attacks not only disrupts system operations and leaks information, but also entails massive economic damage. Recently, cyber attacks have a distinct goal and use advanced attack tools and techniques to accurately infiltrate the target. In order to minimize the damage caused by such an intelligent cyber attack, it is necessary to block the cyber attack at the beginning or during the attack to prevent it from invading the target's core system. Recently, technologies for predicting cyber attack paths and analyzing risk level of cyber attack using big data or artificial intelligence technologies are being studied. In this paper, a cyber attack path analysis method using attack tree and RFI is proposed as a basic algorithm for the development of an automated cyber attack path prediction system. The attack path is visualized using the attack tree, and the priority of the path that can move to the next step is determined using the RFI technique in each attack step. Based on the proposed mechanism, it can contribute to the development of an automated cyber attack path prediction system using big data and deep learning technology.