For a graph G, G → (a1, a2, · · ·, ar)v means that in every r-coloring of the vertices in G, there exists a monochromatic ai-clique of color i for some i∈{1, 2, · · ·, r}. The vertex Fo...
详细信息
A novel image threshold selection approach based on structural similarity (SSIM) is proposed. The thresholded image is obtained first, then comparison regions are extracted based on the local variance of the neighborh...
详细信息
A novel image threshold selection approach based on structural similarity (SSIM) is proposed. The thresholded image is obtained first, then comparison regions are extracted based on the local variance of the neighborhood of the thresholded image. Due to the characteristic of comparison regions, the conventional SSIM expression is simplified as a nonparametric form, and the partial SSIM (PSSIM) is defined. The optimal threshold is selected by maximizing the PSSIM criterion function at last. Besides the introduction of a novel approach, this is also the first attempt to expand the application scope of SSIM to range image thresholding in general. The proposed approach has an advantage over thresholding methods based on the histogram. The method was tested on a variety of images including the synthetic image and real images. Experimental results show that the proposed approach achieves better applicability, preferable ability for extracting object and better anti-noise capability than popular methods.
作者：
Z. ChenJ. G. LiuG. Y. WangIntelligence Control
Institute for Pattern Recognition and Artificial Intelligence and Multi-Spectral Information Processing State Key Laboratory Huazhong University of Science and Technology China Intelligence Control
Institute for Pattern Recognition and Artificial Intelligence and Multi-Spectral Information Processing State Key Laboratory Huazhong University of Science and Technology China
According to the drawback of the traditional circle target extraction algorithm from high resolution remote sensing imagery used by Hough Transform, such as computation complexity, low efficiency and etc, a new circle...
详细信息
According to the drawback of the traditional circle target extraction algorithm from high resolution remote sensing imagery used by Hough Transform, such as computation complexity, low efficiency and etc, a new circle target extraction method is proposed in this paper which can extract multiple circle targets with different radius at one time. First, the Average Absolute Difference is implemented to enhance the edge of the circle targets and suppress the noise of the background. Secondly, the locally self-adaptive segmentation algorithm is implemented to obtain the binary image. Thirdly, the thinning algorithm based on model computation is implanted to obtain the single pixel edge of the circle targets and in order to reduce the computation times in the following process. Furthermore, a pruning algorithm is necessary; finally, a modified Hough transform algorithm is proposed to obtain the circle targets. The experimental results demonstrate that the new circle targets algorithm can extract the multiple circle targets quickly and accurately, which has three advantages: low time consuming, high detection rate, robust to noise and fragmentary boundaries.
Gender is an important demographic attribute of human beings, automatic face based gender classification has promising applications in various fields. Previous methods mainly deal with frontal face images, which in ma...
详细信息
ISBN:
(纸本)9781457713583
Gender is an important demographic attribute of human beings, automatic face based gender classification has promising applications in various fields. Previous methods mainly deal with frontal face images, which in many cases can not be easily obtained. In contrast, we concentrate on gender classification based on face profiles and ear images in this paper. Hierarchical and discriminative bag of features technique is proposed to extract powerful features which are classified by support vector classification (SVC) with histogram intersection kernel. With the output of SVC, fusion of multi-modalities is performed at the score level based on Bayesian analysis to improve the accuracy. Experiments are conducted using texture images of the UND biometrics data sets Collection F, and average classification accuracy of 97.65% is achieved, which is comparable to the state of the art. Our work can be used in cooperate with existing frontal face based methods for accurate multi- view gender classification.
This paper presents a novel method for synthesizing artificial visual light (VIS) face images from near-infrared (NIR) inputs. Active NIR imaging is now widely employed because it is unobtrusive, invariant of environm...
详细信息
This paper presents a novel method for synthesizing artificial visual light (VIS) face images from near-infrared (NIR) inputs. Active NIR imaging is now widely employed because it is unobtrusive, invariant of environmental illuminations, and can penetrate glasses and sweats. Unfortunately, NIR imaging exhibits discrepant photic properties compared with VIS imaging. Based on recent results of re search on compressive sensing, natural images can be compressed and recovered with an overcomplete dictionary by sparse representation coefficients. In our approach a pair wise dictionary is trained from randomly sampled coupled face patches, which contains sparse coded base functions to reconstruct representation coefficients via l 1 -minimization. We will demonstrate that this method is robust to moderate pose and expression variations, and is efficient in computing. Comparative experiments are conducted with state-of the-art algol 1 -minimization. We will demonstraterithms.
This paper presents a novel fast single-pass contour tracing algorithm in a binary image. The proposed algorithm is viewed as follow steps: firstly a set of contour segments of all object contours can be generated and...
详细信息
ISBN:
(纸本)9781612847719
This paper presents a novel fast single-pass contour tracing algorithm in a binary image. The proposed algorithm is viewed as follow steps: firstly a set of contour segments of all object contours can be generated and traced in a top-down line scan fashion; then all contour segments are employed to be integrated into respective intact contours; finally all results are converted into the chain code as the final output. This algorithm can extract multiple contours of an image in one pass and never lose any outer and inner contour of object region. It is faster on implementation. Experiments results prove those advantages.
According to the feature that the gray distribution of the transition region (locating between the objects and the background) is more scattered than that of the regions of targets or background in an image, this pape...
详细信息
According to the feature that the gray distribution of the transition region (locating between the objects and the background) is more scattered than that of the regions of targets or background in an image, this paper proposes a novel method concerning about transition region extraction and segmentation, which is based on local variance of different areas of an image. The experiments indicate that the proposed approach can achieve better segmentation results than the local complexity method (one of the previous methods for extracting transition region).What's more, the novel approach outperforms the local complexity method about more complete and more accurate transition regions, less interference from backgrounds, more detail information of segmented targets, clearer and better segmented targets, easier calculation, and higher processing speed.
In this paper, we denote a color image by a quaternion function, then find edge points by solving the maximum of quaternion fractional directional differentiation(QFDD)'s norm. This method is called edge detection...
详细信息
In this paper, we denote a color image by a quaternion function, then find edge points by solving the maximum of quaternion fractional directional differentiation(QFDD)'s norm. This method is called edge detection based on QFDD. Experiments indicate that the method has special advantages. Comparing with Canny, LOG, Sobel, and general fractional differentiation, we discover that QFDD has fewer false negatives in the textured regions and is also better at detecting edges which are partially defined by texture, which means we will obtain better results in the interesting regions by QFDD and these results are more consistent with the characteristics of human visual system.
SIFT (Scale Invariant Feature Transform) is one of most popular approach for feature detection and matching. Many parallelized algorithms have been proposed to accelerate SIFT to apply into real-time systems. This pap...
详细信息
SIFT (Scale Invariant Feature Transform) is one of most popular approach for feature detection and matching. Many parallelized algorithms have been proposed to accelerate SIFT to apply into real-time systems. This paper divides the researches into three different categories, that is, optimizing parallel algorithms based on general purpose multi-core processors, designing customized multi-core processor dedicated for SIFT and implementing SIFT based FPGA (Field Programmable Gate Arrays). Overview of the three type researches and analysis of task-level parallelism are presented in this paper.
By using wavelet transform modulus maximum principle for non-stationary signal singularity detection is a kind of very good method. Through to the various wavelet singularity extracted, the analysis results can be div...
详细信息
By using wavelet transform modulus maximum principle for non-stationary signal singularity detection is a kind of very good method. Through to the various wavelet singularity extracted, the analysis results can be divided into four types: accurate location, the approximate location, overlapping effect, rim effect. According to the classification we learn the optimal wavelet basis should has the following features: the optimum wavelet basis should have strong ability of detecting and precision, and at the same time the influence of overlap and rim should be as small as possible. According to these characteristics, a discriminant function is constructed. The wavelet basis makes the largest discriminant function value is optimal. The experimental results show that the method in this paper according to find out the optimum wavelet basis did more than other wavelet detection effect better.
暂无评论