• KSII Transactions on Internet and Information Systems (TIIS)
• /
• 제8권6호
• /
• pp.1964-1981
• /
• 2014

An improved Mean-Shift (MS) tracker called joint CB-LBWH, which uses a combined weighted-histogram scheme of CBWH (Corrected Background-Weighted Histogram) and LBWH (likelihood-based Background-Weighted Histogram), is presented. Joint CB-LBWH is based on the notion that target representation employs both feature saliency and confidence to form a compound weighted histogram criterion. As the more prominent and confident features mean more significant for tracking the target, the tuned histogram by joint CB-LBWH can reduce the interference of background in target localization effectively. Comparative experimental results show that the proposed joint CB-LBWH scheme can significantly improve the efficiency and robustness of MS tracker when heavy occlusions and complex scenes exist.

• KSII Transactions on Internet and Information Systems (TIIS)
• /
• 제10권3호
• /
• pp.1377-1389
• /
• 2016

This paper presents a target model update scheme for the mean-shift tracking with background weighted histogram. In the scheme, the target candidate histogram is corrected by considering the back-projection weight of each pixel in the kernel after the best target candidate in the current frame image is chosen. In each frame, the target model is updated by the weighted average of the current target model and the corrected target candidate. We compared our target model update scheme with the previous ones by applying several test sequences. The experimental results showed that the object tracking accuracy was greatly improved by using the proposed scheme.

• Journal of Multimedia Information System
• /
• 제2권4호
• /
• pp.303-310
• /
• 2015

Vulnerable road users such as bike, motorcycle, small automobiles, and etc. are easily attacked or threatened with bigger vehicles than them. So this paper suggests a new approach two-wheelers detection system riding on people based on modified histogram of oriented gradients (HOGs) which is weighted by curvature and local correlation coefficient. This correlation coefficient between two variables, in which one is the person riding a bike and other is its background, can represent correlation relation. First, we extract edge vectors using the curvature of Gaussian and Histogram of Oriented Gradients (HOG) which includes gradient information and differential magnitude as cell based. And then, the value, which is calculated by the correlation coefficient between the area of each cell and one of bike, can be used as the weighting factor in process for normalizing the HOG cell. This paper applied the Adaboost algorithm to make a strong classification from weak classification. The experimental results validate the effectiveness of our proposed algorithm show higher than that of the traditional method and under challenging, such as various two-wheeler postures, complex background, and even conclusion.

• Journal of Electrical Engineering and Technology
• /
• 제10권4호
• /
• pp.1877-1885
• /
• 2015

This paper presents a new Nearest-Neighbors based weighted representation for images and weighted K-Nearest-Neighbors (WKNN) classifier to improve the precision of image classification using the Bag of Visual Words (BOVW) based models. Scale-invariant feature transform (SIFT) features are firstly extracted from images. Then, the K-means++ algorithm is adopted in place of the conventional K-means algorithm to generate a more effective visual dictionary. Furthermore, the histogram of visual words becomes more expressive by utilizing the proposed weighted vector quantization (WVQ). Finally, WKNN classifier is applied to enhance the properties of the classification task between images in which similar levels of background noise are present. Average precision and absolute change degree are calculated to assess the classification performance and the stability of K-means++ algorithm, respectively. Experimental results on three diverse datasets: Caltech-101, Caltech-256 and PASCAL VOC 2011 show that the proposed WVQ method and WKNN method further improve the performance of classification.

• 한국항행학회논문지
• /
• 제15권6호
• /
• pp.1132-1137
• /
• 2011

본 논문에서는 사람의 신체 일부분을 추적하는 시스템을 위해서 피부영역을 추출하고 여러 개의 영역을 추적하는 다중 CAMShift 알고리즘(Multi Continuously Adaptive Mean Shift Algorithm)을 제안하였다. 입력 영상에서 피부영역을 추출하기 위해 영상의 RGB의 특정값을 기준으로 피부색에 적응적인 임계값을 적용하였다. 이때 적용된 피부영역을 양손, 얼굴 등에 초기 윈도우를 설정하였다. 이 영역들을 추적함에 있어 영역들 사이에 폐색 영역을 회피하기 위해 가우시안 배경 모델(Gaussian Background Model)을 사용하여 각 추적 영역들을 제한하였다. 또한 폐색영역에 가중치를 부가하여 확률분포영상에서 중심값을 이동시켜 폐색 영역을 회피하였다. 실험 결과 다중 물체들에 강인한 추적을 보이고 유사한 색상을 갖는 물체의 폐색 시에도 우수한 결과를 보임을 확인하였다.

• 한국산학기술학회논문지
• /
• 제12권8호
• /
• pp.3670-3676
• /
• 2011

셀 애니메이션은 배경이 하나의 셀로 표현되고, 장면이 변화될 경우에는 배경이 변경되기 때문에 장면전환시 비교적 큰 변화가 일어난다. 그리고 실제로 카메라를 이용하여 촬영한 영상과는 달리 사람이 직접 그리다 보니 사용된 색상의 종류 또한 그렇게 많지 않다. 본 논문에서는 애니메이션의 이러한 특성을 최대한 반영하고 보다 효과적으로 셀 애니메이션의 장면전환을 검출하기 위해서 색상과 블록 단위의 히스토그램을 단계적으로 활용하는 새로운 애니메이션의 장면전환 검출 기법을 제안한다. 제안된 알고리즘은 연속적으로 입력되는 애니메이션 영상을 받아들인 후 먼저 칼라공간을 HSI 공간으로 변형하고, 두 영상 사이의 색상 값의 차연산을 수행하여 인접한 영상이 장면전환 후보인지를 1차적으로 판단한다. 만일, 인접한 영상이 장면전환 후보군으로 판단되면 부 영역별로 색상 히스토그램을 작성하고, 여기에 가중치를 적용하여 장면전환이 발생했는지의 유무를 최종적으로 판단한다. 본 논문의 실험에서는 제안된 애니메이션의 장면전환 검출 방법이 기존의 장면전환 검출 방법에 비해 보다 우수하다는 것을 보인다.