• 한국멀티미디어학회논문지
• /
• 제20권5호
• /
• pp.769-781
• /
• 2017

Deep learning networks like Convolutional Neural Networks (CNNs) show successful performances in many computer vision applications such as image classification, object detection, and so on. For implementation of deep learning networks in embedded system with limited processing power and memory, deep learning network may need to be simplified. However, simplified deep learning network cannot learn every possible scene. One realistic strategy for embedded deep learning network is to construct a simplified deep learning network model optimized for the scene images of the installation place. Then, automatic training will be necessitated for commercialization. In this paper, as an intermediate step toward automatic training under fisheye camera environments, we study more precise human localization in fisheye images, and propose an accurate human localization method, Automatic Ground-Truth Labelling Method (AGTLM). AGTLM first localizes candidate human object bounding boxes by utilizing GoogLeNet-LSTM approach, and after reassurance process by GoogLeNet-based CNN network, finally refines them more correctly and precisely(tightly) by applying saliency object detection technique. The performance improvement of the proposed human localization method, AGTLM with respect to accuracy and tightness is shown through several experiments.

• Imaging Science in Dentistry
• /
• 제49권1호
• /
• pp.1-7
• /
• 2019

Purpose: Artificial intelligence (AI), represented by deep learning, can be used for real-life problems and is applied across all sectors of society including medical and dental field. The purpose of this study is to review articles about deep learning that were applied to the field of oral and maxillofacial radiology. Materials and Methods: A systematic review was performed using Pubmed, Scopus, and IEEE explore databases to identify articles using deep learning in English literature. The variables from 25 articles included network architecture, number of training data, evaluation result, pros and cons, study object and imaging modality. Results: Convolutional Neural network (CNN) was used as a main network component. The number of published paper and training datasets tended to increase, dealing with various field of dentistry. Conclusion: Dental public datasets need to be constructed and data standardization is necessary for clinical application of deep learning in dental field.

• 한국멀티미디어학회논문지
• /
• 제24권3호
• /
• pp.382-395
• /
• 2021

In this study, character recognition using deep learning is performed among the various defects in the PCB, the purpose of which is to check whether the printed characters are printed correctly on top of components, or the incorrect parts are attached. Generally, character recognition may be perceived as not a difficult problem when considering MNIST, but the printed letters on the PCB component data are difficult to collect, and have very high redundancy. So if a deep learning model is trained with original data without any preprocessing, it can lead to over fitting problems. Therefore, this study aims to reduce the redundancy to the smallest dataset that can represent large amounts of data collected in limited production sites, and to create datasets through data enhancement to train a flexible deep learning model can be used in various production sites. Moreover, ResNet model verifies to determine which combination of datasets is the most effective. This study discusses how to reduce and augment data that is constantly occurring in real PCB production lines, and discusses how to select coresets to learn and apply deep learning models in real sites.

• Wind and Structures
• /
• 제34권6호
• /
• pp.511-523
• /
• 2022

Detecting the icing on wind turbine blades built-in cold regions with conventional methods is always a very laborious, expensive and very difficult task. Regarding this issue, the use of smart systems has recently come to the agenda. It is quite possible to eliminate this issue by using the deep learning method, which is one of these methods. In this study, an application has been implemented that can detect icing on wind turbine blades images with visualization techniques based on deep learning using images. Pre-trained models of Resnet-50, VGG-16, VGG-19 and Inception-V3, which are well-known deep learning approaches, are used to classify objects automatically. Grad-CAM, Grad-CAM++, and Score-CAM visualization techniques were considered depending on the deep learning methods used to predict the location of icing regions on the wind turbine blades accurately. It was clearly shown that the best visualization technique for localization is Score-CAM. Finally, visualization performance analyses in various cases which are close-up and remote photos of a wind turbine, density of icing and light were carried out using Score-CAM for Resnet-50. As a result, it is understood that these methods can detect icing occurring on the wind turbine with acceptable high accuracy.

• Investigative Magnetic Resonance Imaging
• /
• 제25권4호
• /
• pp.293-299
• /
• 2021

Purpose: To accelerate magnetic resonance fingerprinting (MRF) by developing a flexible deep learning reconstruction method. Materials and Methods: Synthetic data were used to train a deep learning model. The trained model was then applied to MRF for different organs and diseases. Iterative reconstruction was performed outside the deep learning model, allowing a changeable encoding matrix, i.e., with flexibility of choice for image resolution, radiofrequency coil, k-space trajectory, and undersampling mask. In vivo experiments were performed on normal brain and prostate cancer volunteers to demonstrate the model performance and generalizability. Results: In 400-dynamics brain MRF, direct nonuniform Fourier transform caused a slight increase of random fluctuations on the T2 map. These fluctuations were reduced with the proposed method. In prostate MRF, the proposed method suppressed fluctuations on both T1 and T2 maps. Conclusion: The deep learning and iterative MRF reconstruction method described in this study was flexible with different acquisition settings such as radiofrequency coils. It is generalizable for different in vivo applications.

• Journal of Information Technology Applications and Management
• /
• 제29권3호
• /
• pp.43-55
• /
• 2022

In this study, with the goal of developing a deep learning-based product recommendation model for effective matching of influencers and products, a deep learning model with a collaborative filtering model combined with generalized matrix decomposition(GMF), a collaborative filtering model based on multi-layer perceptron (MLP), and neural collaborative filtering and generalized matrix Factorization (NeuMF), a hybrid model combining GMP and MLP was developed and tested. In particular, we utilize one-class problem free boosting (OCF-B) method to solve the one-class problem that occurs when training is performed only on positive cases using implicit feedback in the deep learning-based collaborative filtering recommendation model. In relation to model selection based on overall experimental results, the MLP model showed highest performance with weighted average precision, weighted average recall, and f1 score were 0.85 in the model (n=3,000, term=15). This study is meaningful in practice as it attempted to commercialize a deep learning-based recommendation system where influencer's promotion data is being accumulated, pactical personalized recommendation service is not yet commercially applied yet.

• Journal of Korean Dental Science
• /
• 제16권2호
• /
• pp.156-163
• /
• 2023

Purpose: The aim of this study was to evaluate the feasibility of deep learning techniques to classify the morphology and severity of peri-implantitis bone defects based on periapical radiographs. Materials and Methods: Based on a pre-trained and fine-tuned ResNet-50 deep learning algorithm, the morphology and severity of peri-implantitis bone defects on periapical radiographs were classified into six groups (class I/II and slight/moderate/severe). Accuracy, precision, recall, and F1 scores were calculated to measure accuracy. Result: A total of 971 dental images were included in this study. Deep-learning-based classification achieved an accuracy of 86.0% with precision, recall, and F1 score values of 84.45%, 81.22%, and 82.80%, respectively. Class II and moderate groups had the highest F1 scores (92.23%), whereas class I and severe groups had the lowest F1 scores (69.33%). Conclusion: The artificial intelligence-based deep learning technique is promising for classifying the morphology and severity of peri-implantitis. However, further studies are required to validate their feasibility in clinical practice.

• International Journal of Computer Science & Network Security
• /
• 제23권10호
• /
• pp.89-96
• /
• 2023

Intrusion detection has been widely studied in both industry and academia, but cybersecurity analysts always want more accuracy and global threat analysis to secure their systems in cyberspace. Big data represent the great challenge of intrusion detection systems, making it hard to monitor and analyze this large volume of data using traditional techniques. Recently, deep learning has been emerged as a new approach which enables the use of Big Data with a low training time and high accuracy rate. In this paper, we propose an approach of an IDS based on cloud computing and the integration of big data and deep learning techniques to detect different attacks as early as possible. To demonstrate the efficacy of this system, we implement the proposed system within Microsoft Azure Cloud, as it provides both processing power and storage capabilities, using a convolutional neural network (CNN-IDS) with the distributed computing environment Apache Spark, integrated with Keras Deep Learning Library. We study the performance of the model in two categories of classification (binary and multiclass) using CSE-CIC-IDS2018 dataset. Our system showed a great performance due to the integration of deep learning technique and Apache Spark engine.

• 대한임베디드공학회논문지
• /
• 제18권6호
• /
• pp.319-325
• /
• 2023

In this study, we proposed a method for hyperparameter optimization in the building and training of a deep learning model designed to process point cloud data collected by a millimeter-wave radar system. The primary aim of this study is to facilitate the deployment of a baseline model in resource-constrained IoT devices. We evaluated a RadHAR baseline deep learning model trained on a public dataset composed of point clouds representing five distinct human activities. Additionally, we introduced a coarse-to-fine hyperparameter optimization procedure, showing substantial potential to enhance model efficiency without compromising predictive performance. Experimental results show the feasibility of significantly reducing model size without adversely impacting performance. Specifically, the optimized model demonstrated a 3.3% improvement in classification accuracy despite a 16.8% reduction in number of parameters compared th the baseline model. In conclusion, this research offers valuable insights for the development of deep learning models for resource-constrained IoT devices, underscoring the potential of hyperparameter optimization and model size reduction strategies. This work contributes to enhancing the practicality and usability of deep learning models in real-world environments, where high levels of accuracy and efficiency in data processing and classification tasks are required.

• 한국군사과학기술학회지
• /
• 제27권2호
• /
• pp.147-155
• /
• 2024

Based on the recently developed deep learning technology, many studies have been conducted on deep learning networks that simultaneously detect and classify targets of interest in synthetic aperture radar(SAR) images. Although numerous research results have been derived mainly with the open SAR ship datasets, there is a lack of work carried out on the deep learning network aimed at detecting and classifying SAR ground targets and trained with the synthetic dataset generated from electromagnetic scattering simulations. In this respect, this paper presents the deep learning network trained with the synthetic dataset and applies it to detecting and classifying real SAR ground targets. With experiment results, this paper also analyzes the network performance according to the composition ratio between the real measured data and the synthetic data involved in network training. Finally, the summary and limitations are discussed to give information on the future research direction.