• Journal of Information Processing Systems
• /
• 제17권2호
• /
• pp.227-241
• /
• 2021

Object detection techniques based on deep learning such as YOLO have high detection performance and precision in a single channel video stream. In order to expand to multiple channel object detection in real-time, however, high-performance hardware is required. In this paper, we propose a novel back-end server framework, a real-time AI vision platform (RAVIP), which can extend the object detection function from single channel to simultaneous multi-channels, which can work well even in low-end server hardware. RAVIP assembles appropriate component modules from the RODEM (real-time object detection module) Base to create per-channel instances for each channel, enabling efficient parallelization of object detection instances on limited hardware resources through continuous monitoring with respect to resource utilization. Through practical experiments, RAVIP shows that it is possible to optimize CPU, GPU, and memory utilization while performing object detection service in a multi-channel situation. In addition, it has been proven that RAVIP can provide object detection services with 25 FPS for all 16 channels at the same time.

• International Journal of Internet, Broadcasting and Communication
• /
• 제11권4호
• /
• pp.76-85
• /
• 2019

We introduce a vision-based object detection method for real-time video surveillance system in low-end edge computing environments. Recently, the accuracy of object detection has been improved due to the performance of approaches based on deep learning algorithm such as Region Convolutional Neural Network(R-CNN) which has two stage for inferencing. On the other hand, one stage detection algorithms such as single-shot detection (SSD) and you only look once (YOLO) have been developed at the expense of some accuracy and can be used for real-time systems. However, high-performance hardware such as General-Purpose computing on Graphics Processing Unit(GPGPU) is required to still achieve excellent object detection performance and speed. To address hardware requirement that is burdensome to low-end edge computing environments, We propose sub-frame analysis method for the object detection. In specific, We divide a whole image frame into smaller ones then inference them on Convolutional Neural Network (CNN) based image detection network, which is much faster than conventional network designed forfull frame image. We reduced its computationalrequirementsignificantly without losing throughput and object detection accuracy with the proposed method.

• 한국방송∙미디어공학회:학술대회논문집
• /
• 한국방송공학회 2009년도 IWAIT
• /
• pp.71-75
• /
• 2009

In this paper, we present a real-time method to detect moving objects in a rotating and zooming camera. It is useful for camera surveillance of fixed but rotating camera, camera on moving car, and so on. We first compensate the global motion, and then exploit the displaced frame difference (DFD) to find the block-wise boundary. For robust detection, we propose a kind of image to combine the detections from consecutive frames. We use the block-wise detection to achieve the real-time speed, except the pixel-wise DFD. In addition, a fast block-matching algorithm is proposed to obtain local motions and then global affine motion. In the experimental results, we demonstrate that our proposed algorithm can handle the real-time detection of common object, small object, multiple objects, the objects in low-contrast environment, and the object in zooming camera.

• 한국산업정보학회:학술대회논문집
• /
• 한국산업정보학회 2003년도 추계공동학술대회
• /
• pp.145-156
• /
• 2003

This paper for real time object tracking in this treatise detect histogram analysis that is accumulation value of binary conversion density and edge information and body that move by real time use of difference Image techniques and proposed method to object tracking. Firstly, we extract edge that can reduce quantity of data keeping information about form of input image in object detection. Object is extracted by performing difference image and binarization in edge image. Area of detected object is determined by threshold value that divide sum of horizontal accumulation value about binary conversion density by value that add horizontalityㆍverticality maximum accumulation value. Object is tracked by comparing similarity with object that is detected in previous frame and present frame. As experiment result, proposed algorithm could improve the object detection speed, and could track object by real time and could track local movement.

• 한국컴퓨터정보학회논문지
• /
• 제27권1호
• /
• pp.1-8
• /
• 2022

본 논문에서는 실시간 객체 탐지(Real-time Object Detection)가 가능한 YOLOv4 모델과 DeepSORT 알고리즘을 활용한 객체 추적(Object Tracking) 기술을 활용하여 CCTV 영상 이미지 기반의 화재 탐지 시스템을 제안한다. 화재 탐지 모델은 10800장의 학습용 데이터로부터 학습되었으며 1000장의 별도 테스트 셋을 통해 검증되었다. 이후 DeepSORT 알고리즘을 통해 탐지된 화재 영역을 추적하여 단일 이미지 내의 화재 탐지율과 영상 내에서의 화재 탐지 유지성능을 증가시켰다. 영상 내의 한 프레임 혹은 단일 이미지에 대한 화재 탐지 속도는 장당 0.1초 이내로 실시간 탐지가 가능함을 확인하였으며 본 논문의 AI 화재 탐지 시스템은 기존의 화재 사고 탐지 시스템 보다 안정적이고 빠른 성능을 지니고 있어 화재현장에 적용 시 화재를 조기 발견하여 빠른 대처 및 발화단계에서의 진화가 가능할 것으로 예상된다.

• 한국정보처리학회:학술대회논문집
• /
• 한국정보처리학회 2021년도 추계학술발표대회
• /
• pp.723-726
• /
• 2021

현재 단일 이미지에서 Object Detection 성능은 매우 좋은 편이다. 하지만 동영상에서는 처리 속도가 너무 느리고 임베디드 시스템에서는 real-time이 힘든 상황이다. 연구 논문에서는 하이엔드 GPU에서 다른 기능 없이 YOLO만 구동했을 때 real-time이 가능하다고 하지만 실제 사용자들은 상대적으로 낮은 사양의 GPU를 사용하거나 CPU를 사용하기 때문에 일반적으로는 자연스러운 real-time을 하기가 힘들다. 본 논문에서는 이러한 제한점을 해결하고자 계산량이 많은 Object Detection model 사용을 줄이는 방안은 제시하였다. 현재 Video영상에서 Object Detection을 수행할 때 매 frame마다 YOLO모델을 구동하는 것에서 YOLO 사용을 줄임으로써 계산 효율을 높였다. 본 논문의 알고리즘은 카메라가 움직이거나 배경이 바뀌는 상황에서도 사용이 가능하다. 속도는 최소2배에서 ~10배이상까지 개선되었다.

• KSII Transactions on Internet and Information Systems (TIIS)
• /
• 제11권8호
• /
• pp.4120-4132
• /
• 2017

In this research, we implement a deformable object simulation system using OpenGL's shader language, GLSL4.3. Deformable object simulation is implemented by using volumetric mass-spring system suitable for real-time simulation among the methods of deformable object simulation. The compute shader in GLSL 4.3 which helps to access the GPU resources, is used to parallelize the operations of existing deformable object simulation systems. The proposed system is implemented using a compute shader for parallel processing and it includes a bounding box-based collision detection solution. In general, the collision detection is one of severe computing bottlenecks in simulation of multiple deformable objects. In order to validate an efficiency of the system, we performed the experiments using the 3D volumetric objects. We compared the performance of multiple deformable object simulations between CPU and GPU to analyze the effectiveness of parallel processing using GLSL. Moreover, we measured the computation time of bounding box-based collision detection to show that collision detection can be processed in real-time. The experiments using 3D volumetric models with 10K faces showed the GPU-based parallel simulation improves performance by 98% over the CPU-based simulation, and the overall steps including collision detection and rendering could be processed in real-time frame rate of 218.11 FPS.

• International journal of advanced smart convergence
• /
• 제11권2호
• /
• pp.22-29
• /
• 2022

In this paper, for the purpose of designing an real-time unmanned monitoring system, the YOLOv5s (small) object detection model was applied on the NVIDIA TX2^TM AI (Artificial Intelligence) edge computing platform in order to design the fundamental function of an unmanned monitoring system that can detect objects in real time. YOLOv5s was applied to the our real-time unmanned monitoring system based on the performance evaluation of object detection algorithms (for example, R-CNN, SSD, RetinaNet, and YOLOv5). In addition, the performance of the four YOLOv5 models (small, medium, large, and xlarge) was compared and evaluated. Furthermore, based on these results, the YOLOv5s model suitable for the design purpose of this paper was ported to the NVIDIA TX2^TM AI edge computing system and it was confirmed that it operates normally. The real-time unmanned monitoring system designed as a result of the research can be applied to various application fields such as an security or monitoring system. Future research is to apply NMS (Non-Maximum Suppression) modification, model reconstruction, and parallel processing programming techniques using CUDA (Compute Unified Device Architecture) for the improvement of object detection speed and performance.

• 한국산학기술학회논문지
• /
• 제19권3호
• /
• pp.693-698
• /
• 2018

본 논문은 최근 물체탐지 분야에서 실시간 물체 탐지 알고리즘으로 주목을 받고 있는 YOLOv2(You Only Look Once) 알고리즘을 이용하여 밀집 영역에 주차되어 있는 자동차 탐지 방법을 제안한다. YOLO의 컨볼루션 네트워크는 전체 이미지에서 한 번의 평가를 통해서 직접적으로 경계박스들을 예측하고 각 클래스의 확률을 계산하고 물체 탐지 과정이 단일 네트워크이기 때문에 탐지 성능이 최적화 되며 빠르다는 장점을 가지고 있다. 기존의 슬라이딩 윈도우 접근법과 R-CNN 계열의 탐지 방법은 region proposal 방법을 사용하여 이미지 안에 가능성이 많은 경계박스를 생성하고 각 요소들을 따로 학습하기 때문에 최적화 및 실시간 적용에 어려움을 가지고 있다. 제안하는 연구는 YOLOv2 알고리즘을 적용하여 기존의 알고리즘이 가지고 있는 물체 탐지의 실시간 처리 문제점을 해결하여 실시간으로 지상에 있는 자동차를 탐지하는 방법을 제안한다. 제안하는 연구 방법의 실험을 위하여 오픈소스로 제공되는 Darknet을 사용하였으며 GTX-1080ti 4개를 탑재한 Deep learning 서버를 이용하여 실험하였다. 실험결과 YOLO를 활용한 자동차 탐지 방법은 기존의 알고리즘 보다 물체탐지에 대한 오버헤드를 감소 할 수 있었으며 실시간으로 지상에 존재하는 자동차를 탐지할 수 있었다.

• 한국자동차공학회논문집
• /
• 제24권2호
• /
• pp.127-136
• /
• 2016

The camera based object detection systems should satisfy the recognition performance as well as real-time constraints. Particularly, in safety-critical systems such as Autonomous Emergency Braking (AEB), the real-time constraints significantly affects the system performance. Recently, multi-core processors and system-on-chip technologies are widely used to accelerate the object detection algorithm by distributing computational loads. However, due to the advanced hardware, the complexity of system architecture is increased even though additional hardwares improve the real-time performance. The increased complexity also cause difficulty in migration of existing algorithms and development of new algorithms. In this paper, to improve real-time performance and design complexity, a task scheduling strategy is proposed for visual object tracking systems. The real-time performance of the vision algorithm is increased by applying pipelining to task scheduling in a multi-core processor. Finally, the proposed task scheduling algorithm is applied to crosswalk detection and tracking system to prove the effectiveness of the proposed strategy.