• Smart Media Journal
• /
• v.9 no.2
• /
• pp.92-98
• /
• 2020

This paper proposes a vehicle license plate recognition system using light weight deep learning models without high-end server. The proposed license plate recognition system consists of 3 steps: [license plate detection]-[character area segmentation]-[character recognition]. SSD-Mobilenet was used for license plate detection, ResNet with localization was used for character area segmentation, ResNet was used for character recognition. Experiemnts using Samsung Galaxy S7 and LG Q9, accuracy showed 85.3% accuracy and around 1.1 second running time.

• Proceedings of the Korea Information Processing Society Conference
• /
• 2023.05a
• /
• pp.79-80
• /
• 2023

CPU 와 GPU 의 성능이 지속적으로 발전함에 따라 객체 인식 인공지능의 정확도와 추론 속도는 점차 향상되고 있으나 이러한 성능을 Edge Device 와 같은 제한된 환경에서 구현하기에 아직 여러 한계점이 존재한다. 본 논문에서는 여러가지 Edge Device 에서 객체 인식을 위한 경량화 된 모델 중 하나인 SSD-Mobilenet-V2 를 활용하여 결과값을 통해 각 Device 간 경향성을 분석하였다. 본 결과를 바탕으로 다양한 환경에서의 객체인식 인공지능 모델의 성능 개선을 위한 연구에 활용할 수 있다.

• Proceedings of the Korea Information Processing Society Conference
• /
• 2019.05a
• /
• pp.398-400
• /
• 2019

Faced with the increasing demand for image recognition on mobile devices, how to run convolutional neural network (CNN) models on mobile devices with limited computing power and limited storage resources encourages people to study efficient model design. In recent years, many effective architectures have been proposed, such as mobilenet_v1, mobilenet_v2 and PeleeNet. However, in the process of feature selection, all these models neglect some information of shallow features, which reduces the capture of shallow feature location and semantics. In this study, we propose an effective framework based on Feature Pyramid Networks to improve the recognition accuracy of deep and shallow images while guaranteeing the recognition speed of PeleeNet structured images. Compared with PeleeNet, the accuracy of structure recognition on CIFA-10 data set increased by 4.0%.

• Smart Media Journal
• /
• v.11 no.7
• /
• pp.94-103
• /
• 2022

Weeds bring disadvantages to crops since they can damage them, and a clean treatment with less pollution and contamination should be developed. Artificial intelligence gives new hope to agriculture to achieve smart farming. This study delivers an automated weeds growth point detection using deep learning. This study proposes a combination of semantic graphics for generating data annotation and U-Net with pre-trained deep learning as a backbone for locating the growth point of the weeds on the given field scene. The dataset was collected from an actual field. We measured the intersection over union, f1-score, precision, and recall to evaluate our method. Moreover, Mobilenet V2 was chosen as the backbone and compared with Resnet 34. The results showed that the proposed method was accurate enough to detect the growth point and handle the brightness variation. The best performance was achieved by Mobilenet V2 as a backbone with IoU 96.81%, precision 97.77%, recall 98.97%, and f1-score 97.30%.

• KIPS Transactions on Software and Data Engineering
• /
• v.9 no.10
• /
• pp.323-328
• /
• 2020

Recently, new approaches that significantly improve performance in object detection and recognition using deep learning technology have been proposed quickly. Of the various techniques for object detection, especially facial object detection (Faster R-CNN, R-CNN, YOLO, SSD, etc), SSD is superior in accuracy and speed to other techniques. At the same time, multiple object detection networks are also readily available. In this paper, among object detection networks, Mobilenet v2 network is used, models combined with SSDs are trained, and methods for detecting objects at a rate of four times or more than conventional performance are proposed using TensorRT engine, and the performance is verified through experiments. Facial object detector was created as an application to verify the performance of the proposed method, and its behavior and performance were tested in various situations.

• Proceedings of the Korean Society of Computer Information Conference
• /
• 2019.07a
• /
• pp.375-377
• /
• 2019

본 논문에서는 관람객 얼굴 표정 및 제스쳐 인식을 기반으로 인터렉티브 디지털콘텐츠의 문제점 추론 시스템을 제안한다. 관람객이 콘텐츠를 체험하고 다른 장소로 이동하기 전까지의 행동 패턴을 기준으로 삼아 4가지 문제점으로 분류한다. 문제점 분류을 위해 관람객이 콘텐츠 체험과정에서 나타낼 수 있는 얼굴 표정 3가지 종류와 제스쳐 5가지를 구분하였다. 실험에서는 입력된 비디오로부터 얼굴 및 손을 검출하기 위해 Adaboost algorithm을 사용하였고, mobilenet v1을 retraining하여 탐지모델을 생성 후 얼굴 표정 및 제스쳐를 검출했다. 이 연구를 통해 인터렉티브 디지털콘텐츠가 지니고 있는 문제점을 추론하여 향후 콘텐츠 개선과 제작에 사용자 중심 설계가 가능하도록 하고 양질의 콘텐츠 생산을 촉진 시킬 수 있을 것이다.

• Proceedings of the Korea Information Processing Society Conference
• /
• 2022.05a
• /
• pp.420-423
• /
• 2022

최근 딥러닝 기술이 모바일 기기에 활발히 적용됨에 따라 다양한 엣지 디바이스에서 신경망 모델의 추론 성능을 측정하는 것이 중요해지고 있다. 하지만 디바이스 별 환경 구성과 런타임별 모델 변환 방식이 다르기 때문에 이를 실제로 수행하는 것은 많은 시간을 필요로 한다. 따라서 본 논문에서는 이기종 환경을 고려하여 추론 성능을 측정할 수 있는 Docker 기반의 프로그램을 구현하였고, 이를 이용하여 다양한 엣지 디바이스에서 최신 모델들의 추론 성능을 측정하였다. 또한, 본 프로그램으로 확보 가능한 추론시간 데이터 기반 추론 성능 예측 연구의 사전 연구로서, 대표적 경량모델인 MobilenetV1 에 대한 연산자별 프로파일링을 수행하여 추론시간의 변화 양상을 관찰하였다.

• Proceedings of the Korea Information Processing Society Conference
• /
• 2020.05a
• /
• pp.574-576
• /
• 2020

최근에는 딥러닝 기술의 발달로 물체 인식 및 검출에 관한 기술들 또한 발탄하고 있다. 검출에 관한 여러 기법(Faster R-CNN, R-CNN, YOLO, SSD 등) 중 SSD는 다른 기법들과는 다르게 높은 정확도와 빠른 속도가 특징이다. 동시에 여러 detection network들도 쉽게 이용이 가능하다. 본 논문에서는 detection netowork중 Mobilenet V2 network를 이용하여 SSD와 결합해 모델을 훈련하고, TensorRT engine을 이용하여 더 빠른 속도로 검출할 수 있는 방법에 대해 논의한다. 이 방법을 통해 face detector를 만들어 여러 상황에서 쓰일 수 있도록 한다.

• The Journal of Society for e-Business Studies
• /
• v.26 no.1
• /
• pp.93-106
• /
• 2021

Many recent studies on deep learning-based road crack detection have shown significantly more improved performances than previous works using algorithm-based conventional approaches. However, many deep learning-based studies are still focused on classifying the types of cracks. The classification of crack types is highly anticipated in that it can improve the crack detection process, which is currently relying on manual intervention. However, it is essential to calculate the severity of the cracks as well as identifying the type of cracks in actual pavement maintenance planning, but studies related to road crack detection have not progressed enough to automated calculation of the severity of cracks. In order to calculate the severity of the crack, the type of crack and the area of the crack in the image must be identified together. This study deals with a method of using Mobilenet-SSD that is deep learning-based object detection techniques to effectively automate the simultaneous detection of crack types and crack areas. To improve the accuracy of object-detection for road cracks, several experiments were conducted to combine the U-Net for automatic segmentation of input image and object-detection model, and the results were summarized. As a result, image masking with U-Net is able to maximize object-detection performance with 0.9315 mAP value. While referring the results of this study, it is expected that the automation of the crack detection functionality on pave management system can be further enhanced.

• KIPS Transactions on Computer and Communication Systems
• /
• v.11 no.10
• /
• pp.323-332
• /
• 2022

As various services using AI technology are being developed, much attention is being paid to AI service production. Recently, AI technology is acknowledged as one of ICT services, a lot of research is being conducted for general-purpose AI service production. In this paper, I describe the research results in terms of systems for AI service production, focusing on the distribution and production of machine learning models, which are the final steps of general machine learning development procedures. Three different Ubuntu systems were built, and experiments were conducted on the system, using data from 2017 validation COCO dataset in combination of different AI models (RFCN, SSD-Mobilenet) and different communication methods (gRPC, REST) to request and perform AI services through Tensorflow serving. Through various experiments, it was found that the type of AI model has a greater influence on AI service inference time than AI machine communication method, and in the case of object detection AI service, the number and complexity of objects in the image are more affected than the file size of the image to be detected. In addition, it was confirmed that if the AI service is performed remotely rather than locally, even if it is a machine with good performance, it takes more time to infer the AI service than if it is performed locally. Through the results of this study, it is expected that system design suitable for service goals, AI model development, and efficient AI service production will be possible.