• 한국멀티미디어학회논문지
• /
• 제19권12호
• /
• pp.1909-1918
• /
• 2016

This paper presents a unified framework for joint Convolutional Neural Network (CNN) based vehicle detection by leveraging multi-spectral image pairs. With the observation that under challenging environments such as night vision and limited light source, vehicle detection in a single color image can be more tractable by using additional far-infrared (FIR) image, we design joint CNN architecture for both RGB and FIR image pairs. We assume that a score map from joint CNN applied to overall image can be considered as confidence of vehicle existence. To deal with various scale ratios of vehicle candidates, multi-scale images are first generated scaling an image according to possible scale ratio of vehicles. The vehicle candidates are then detected on local maximal on each score maps. The generation of overlapped candidates is prevented with non-maximal suppression on multi-scale score maps. The experimental results show that our framework have superior performance than conventional methods with a joint framework of multi-spectral image pairs reducing false positive generated by conventional vehicle detection framework using only single color image.

• 정보처리학회논문지D
• /
• 제17D권6호
• /
• pp.463-470
• /
• 2010

본 논문에서는 DCT 데이터에서 영상 데이터로의 해독 및 이진화 과정을 생략하고 컬러 영상의 DCT 관련 원자료를 사용하는 방법에 기반을 둔 매우 빠르고 안정적인 문자열 구간 분리 모형을 제안하였다. DCT 블록에 저장된 DC 및 3개의 주요 AC 변수들을 조합하여 축소된 저해상도 회색 영상을 만들고 횡렬 및 종렬 투영법을 통해 얻어진 픽셀 값의 히스토그램을 분석하여 문자 열 구간 사이에 존재하는 백색의 띠 공간을 찾아내었다. 이 과정 중 탐색되지 않은 문자 열 구간은 마코프 모델을 사용하여 숨겨진 주기를 찾아내어 복원하였다. 본 논문에 실험 결과를 제시하였으며 기존의 방법보다 약 40 - 100배 빠른 방법임을 입증하였다.

• 한국정보전자통신기술학회논문지
• /
• 제12권3호
• /
• pp.186-191
• /
• 2019

본 논문은 고속 특징추출 알고리즘의 구현 방법을 제안한다. 제안하는 방법은 블록 유형 분류 알고리즘을 기반으로, 블록 유형 분류 알고리즘 적용 시, 영상 특징 정보가 발생하지 않는 스무스 블록에서 연산을 생략하여 영상 특징 검출에 필요한 연산시간을 감소시킬 수 있다. 200장의 표준 테스트 이미지를 활용해 매크로 블록의 크기를 $64{\times}64$로 나누어 스무스 블록의 발생 빈도를 측정한 결과 전체의 29.5%만큼 발생하는 것을 정량적으로 확인했다. 이 의미는 다양한 영상 정보를 포함하고 있는 표준 테스트 이미지 내에서는 29.5%에 해당하는 만큼 연산의 복잡도를 감소시킬 수 있다는 의미를 나타낸다. 제안된 방법을 케니 윤곽선 검출 알고리즘에 적용하면 이차원 미분 필터, 그라디언트 크기 및 방향 연산, 비최대 억제, 적응형 임계값 연산, 히스테리시스 임계 처리와 같은 총 다섯 단계의 영상처리에 필요한 지연시간을 완전히 제거할 수 있다. 이와 같은 방법으로 다양한 특징 검출 알고리즘에 블록 유형 구분 알고리즘을 적용해, 연산에 필요한 시간을 감소할 수 있을 것을 기대한다.

• KSII Transactions on Internet and Information Systems (TIIS)
• /
• 제16권7호
• /
• pp.2131-2153
• /
• 2022

To achieve accurate detection of tuberculosis (TB) areas in chest radiographs, we design a chest X-ray TB area detection algorithm. The algorithm consists of two stages: the chest X-ray TB classification network (CXTCNet) and the chest X-ray TB area detection network (CXTDNet). CXTCNet is used to judge the presence or absence of TB areas in chest X-ray images, thereby excluding the influence of other lung diseases on the detection of TB areas. It can reduce false positives in the detection network and improve the accuracy of detection results. In CXTCNet, we propose a channel attention mechanism (CAM) module and combine it with DenseNet. This module enables the network to learn more spatial and channel features information about chest X-ray images, thereby improving network performance. CXTDNet is a design based on a sparse object detection algorithm (Sparse R-CNN). A group of fixed learnable proposal boxes and learnable proposal features are using for classification and location. The predictions of the algorithm are output directly without non-maximal suppression post-processing. Furthermore, we use CLAHE to reduce image noise and improve image quality for data preprocessing. Experiments on dataset TBX11K show that the accuracy of the proposed CXTCNet is up to 99.10%, which is better than most current TB classification algorithms. Finally, our proposed chest X-ray TB detection algorithm could achieve AP of 45.35% and AP50 of 74.20%. We also establish a chest X-ray TB dataset with 304 sheets. And experiments on this dataset showed that the accuracy of the diagnosis was comparable to that of radiologists. We hope that our proposed algorithm and established dataset will advance the field of TB detection.

• Molecules and Cells
• /
• 제38권4호
• /
• pp.373-379
• /
• 2015

Pyruvate kinase M2 isoform (PKM2), a rate-limiting enzyme in the final step of glycolysis, is known to be associated with the metabolic rewiring of cancer cells, and considered an important cancer therapeutic target. Herein, we report a novel PKM2 activator, PA-12, which was identified via the molecular docking-based virtual screening. We demonstrate that PA-12 stimulates the pyruvate kinase activity of recombinant PKM2 in vitro, with a half-maximal activity concentration of $4.92{\mu}M$, and effectively suppresses both anchorage-dependent and -independent growth of lung cancer cells in non-essential amino acid-depleted medium. In addition, PA-12 blocked the nuclear translocalization of PKM2 in lung cancer cells, resulting in the inhibition of hypoxia response element (HRE)-mediated reporter activity as well as hypoxia-inducible factor 1 (HIF-1) target gene expression, eventually leading to the suppression of cell viability under hypoxia. We also verified that the effects of PA-12 were dependent on PKM2 expression in cancer cells, demonstrating the specificity of PA-12 for PKM2 protein. Taken together, our data suggest that PA-12 is a novel and potent PKM2 activator that has therapeutic implications for lung cancer.