• Journal of Information Processing Systems
• /
• 제15권1호
• /
• pp.47-54
• /
• 2019

This study is concerned with the mechanism and structure of an optical microscope and an automatic multi-focus algorithm for automatically selecting sharp images from multiple foci of a cell. To obtain precise cell images quickly, a z-axis actuator with a resolution of $0.1{\mu}m$ was designed to control an optical microscope Moreover, a lighting control system was constructed to select the color and brightness of light that best suit the object being viewed. Cell images are captured by the instrument and the sharpness of each image is determined using Gaussian and Laplacian filters. Next, cubic spline interpolation and peak detection algorithms are applied to automatically find the most vivid points among multiple images of a single object. A cancer cell imaging experiment using propidium iodide staining confirmed that a sharp multipoint image can be obtained using this microscope. The proposed system is expected to save time and effort required to extract suitable cell images and increase the convenience of cell analysis.

• 한국정보처리학회:학술대회논문집
• /
• 한국정보처리학회 2016년도 추계학술발표대회
• /
• pp.506-509
• /
• 2016

We present real-time pedestrian detection that exploit accuracy of Faster R-CNN network. Faster R-CNN has shown to success at PASCAL VOC multi-object detection tasks, and their ability to operate on raw pixel input without the need to design special features is very engaging. Therefore, in this work we apply and adjust Faster R-CNN to single object detection, which is pedestrian detection. The drawback of Faster R-CNN is its failure when object size is small. Previously, small sized object problem was solved by Scale-aware Network. We incorporate Scale-aware Network to Faster R-CNN. This made our method Scale-aware Faster R-CNN (DF R-CNN) that is both fast and very accurate. We separated Faster R-CNN networks into two sub-network, that is one for large-size objects and another one for small-size objects. The resulting approach achieves a 28.3% average miss rate on the Caltech Pedestrian detection benchmark, which is competitive with the other best reported results.

• KSII Transactions on Internet and Information Systems (TIIS)
• /
• 제13권8호
• /
• pp.4160-4173
• /
• 2019

Borescopes facilitate the inspection of areas inside machines and systems that are not directly accessible for visual inspection. They offer real-time, up-close access to confined and hard-to-access spaces without having to dismantle or destructure the object under inspection. Borescopes are ideal instruments for routine maintenance, quality inspection and monitoring of systems and structures. The main application being fault or defect detection, it is useful to have measuring capability to quantify object dimensions in a target area. High-end borescopes use multi-optic solutions to provide measurement information of viewed objects. Multi-optic solutions can provide accurate measurements at the expense of structural complexity and cost increase. Measuring functionality is often unavailable in low-end, single camera borescopes. In this paper, a single camera measurement solution that enables the size estimation of viewed objects is proposed. The proposed solution computes and overlays a scaled grid of known spacing value over the screen view, enabling the human inspector to estimate the size of the objects in view. The proposed method provides a simple means of measurement that is applicable to low-end borescopes with no built-in measurement capability.

• 한국컴퓨터정보학회논문지
• /
• 제22권3호
• /
• pp.45-51
• /
• 2017

In this paper, we propose a consecutive-frame super-resolution method to tackle a moving object problem. The super-resolution is a method restoring a high resolution image from a low resolution image. The super-resolution is classified into two types, briefly, single-frame super-resolution and consecutive-frame super-resolution. Typically, the consecutive-frame super-resolution recovers a better than the single-frame super-resolution, because it use more information from consecutive frames. However, the consecutive-frame super-resolution failed to recover the moving object. Therefore, we proposed an improved method via moving object detection. Experimental results showed that the proposed method restored both the moving object and the background properly.

• 반도체디스플레이기술학회지
• /
• 제17권2호
• /
• pp.67-70
• /
• 2018

Object detection and tracking is an exciting and interesting research area in the field of computer vision, and its technologies have been widely used in various application systems such as surveillance, military, and augmented reality. This paper proposes and implements a novel and more robust object recognition and tracking system to localize and track multiple objects from input images, which estimates target state using the likelihoods obtained from multiple CNNs. As the experimental result, the proposed algorithm is effective to handle multi-modal target appearances and other exceptions.

• 스마트미디어저널
• /
• 제12권9호
• /
• pp.45-59
• /
• 2023

Weakly supervised object localization (WSOL) is a task of localizing an object in an image using only image-level labels. Previous studies have followed the conventional class activation mapping (CAM) pipeline. However, we reveal the current CAM approach suffers from problems which cause original CAM could not capture the complete defects features. This work utilizes a convolutional neural network (CNN) pretrained on image-level labels to generate class activation maps in a multi-scale manner to highlight discriminative regions. Additionally, a vision transformer (ViT) pretrained was treated to produce multi-head attention maps as an auxiliary detector. By integrating the CNN-based CAMs and attention maps, our approach localizes defective regions without requiring bounding box or pixel-level supervision during training. We evaluate our approach on a dataset of apple images with only image-level labels of defect categories. Experiments demonstrate our proposed method aligns with several Object Detection models performance, hold a promise for improving localization.

• 대한공간정보학회지
• /
• 제21권3호
• /
• pp.11-17
• /
• 2013

사회 변화에 따라 산지이용 수요가 증가하고 다양화되면서 산림을 훼손하고, 타 용도로 활용하는 산지의 면적이 증가하고 있다. 이에 최근 훼손된 산지의 면적을 효과적으로 확인하기 위하여 두 시기의 항공사진을 활용한 훼손 산지 변화탐지 기법을 연구하였다. 본 연구에서 개발한 기법은 객체기반 변화탐지 형식으로, 영상 혼합 - 객체 분할 - 객체 병합 - 노이즈 제거 - 훼손지 추출의 5가지 단계로 진행되었다. 훼손 산지에 적합한 객체생성 수준을 선정하고, 객체를 분할 병합하는 과정을 통해 객체 간의 관계와 각 객체가 지닌 분광 특성 및 정황적(Contextual) 정보를 활용하여 신규 훼손 산지를 추출하였다. 시범 영역 테스트 결과, 전체 판독범위의 12%에 해당하는 훼손 산지를 추출하였고 육안판독 훼손산지의 평균 96%를 포함함으로써, 육안판독 전 후의 보완 자료로서의 가치와 자동추출의 가능성을 확인하였다.

• 한국정보통신학회논문지
• /
• 제24권11호
• /
• pp.1437-1444
• /
• 2020

컴퓨터 비전 시스템의 발달로 보안, 생체인식, 의료영상, 자율주행 등의 분야에 많은 발전이 있었다. 자율주행 분야에서는 특히 딥러닝을 이용한 객체인식, 탐지 기법이 주로 사용되는데, 자동차가 갈 수 있는 영역을 판단하기 위한 도로영역 인식이 특히 중요한 문제이다. 도로 영역은 일반적인 객체탐지에서 활용되는 사각영역인식과는 달리 비정형적인 형태를 띠므로, ROI 기반의 객체인식 구조는 적용할 수 없다. 본 논문에서는 Semantic segmentation 기법을 사용한 비정형적인 도로영역 인식에 맞는 심층 신경망 구조를 제안한다. 또한 도로영역에 특화된 네트워크 구조인 Multi-scale semantic segmentation 기법을 사용하여 성능이 개선됨을 입증하였다.

• 대한전자공학회논문지SP
• /
• 제47권1호
• /
• pp.9-16
• /
• 2010

이동 객체 검출은 입력 영상에서 배경과 다른 전경 객체를 찾는 것을 말하는 것으로 지능 영상 감시, HCI, 객체 기반 영상 압축 등의 여러 영상 처리 응용 분야에서 필요한 과정이다. 기존의 이동 객체 검출 알고리즘은 상당한 계산량을 요구하여 다채널 영상 감시 응용, 또는 임베디드 시스템에서의 단일 채널의 실시간 응용에 사용하는 데 애로가 많다. 보다 정확한 이동 객체 검출을 위하여 필요한 과정인 전경 마스크 정정은 보통 열림, 닫힘 등의 모폴로지 연산을 통해 수행된다. 모폴로지 연산은 계산량이 적지 않고 게다가 프로세싱 방법이 달라 이동 객체 검출의 다음 단계인 연결 요소 레이블링 루틴과 동시에 처리되기 어렵다. 본 논문에서는 먼저 모폴로지 연산과는 달리 연결 요소 레이블링 루틴에서 사용되는 주변 픽셀 점검 과정을 활용한 전경 마스크 정정 알고리즘인 "주변 전경 픽셀 전파"을 고안하고, 이를 활용하여 전경 마스크 정정과 연결 요소 레이블링이 동시에 수행될 수 있는 이동 객체 검출 방법을 제안한다. 실험을 통해, 제안된 이동 객체 검출 방법이 기존의 모폴로지 연산을 사용한 방법 보다 정확하게 이동 객체를 검출하였으며, 대상 실험 영상 프레임 및 비디오에 대해서는 최소 4배 이상 신속하게 처리됨을 확인하였다.

• 제어로봇시스템학회논문지
• /
• 제13권11호
• /
• pp.1074-1081
• /
• 2007

This paper presents a method for real-time face detection and tracking which combined Adaboost and Camshift algorithm. Adaboost algorithm is a method which selects an important feature called weak classifier among many possible image features by tuning weight of each feature from learning candidates. Even though excellent performance extracting the object, computing time of the algorithm is very high with window size of multi-scale to search image region. So direct application of the method is not easy for real-time tasks such as multi-task OS, robot, and mobile environment. But CAMshift method is an improvement of Mean-shift algorithm for the video streaming environment and track the interesting object at high speed based on hue value of the target region. The detection efficiency of the method is not good for environment of dynamic illumination. We propose a combined method of Adaboost and CAMshift to improve the computing speed with good face detection performance. The method was proved for real image sequences including single and more faces.