Network
Latest external collaboration on country level. Dive into details by clicking on the dots.
Publication
Featured researches published by Hongmin Liu.
international congress on image and signal processing | 2011
Li-Qin Jia; Cheng-zhang Peng; Hongmin Liu; Zhiheng Wang
In this paper, we present a fast randomized circle detection algorithm applied to determine the centers and radii of circular components. Firstly, the gradient of each pixel in the image is computed using Gaussian template. Then, the edge map of the image, obtained by applying canny edge detector, is tackled to acquire the curves consisting of 8-adjacency connected edge points. Subsequently, for the detection of the center, N edge points for each curve are picked up, and the point, passed through by the most gradient lines of the edge points, corresponds to a center. The radius can be received by computing the distance between the center and the corresponding edge points. The algorithm performs much better in terms of efficiency compared to randomized circle detection algorithm (RCD), in which a mass of accumulations are done by random sampling. Synthetic images and natural images are used to test the capability of the proposed algorithm. The experimental results indicate that the presented algorithm consumes less computing resources, has excellent performance for detection of single circle, multiple circles, concentric circles, partial circles and overlapped circles, and also has good accuracy despite the presence of different noises and interference.
international conference on machine learning and cybernetics | 2011
Li-Qin Jia; Hongmin Liu; Zhiheng Wang; Hong Chen
We present a non-HT circle detection algorithm applied to search the centers and radii of circular or partially circular components present in the image. The line coincident with the gradient vector of each edge point and passing through the corresponding edge point is defined first. Then, for every pixel in the image, the number of the lines passing through this pixel is defined as the energy of the pixel. The feature circle energy (FCE) distribution map of the whole image is therefore obtained and the local maxima are corresponding to the centers of potential circles. For the detection of radius, the gradient magnitudes in assigned region are accumulated and its variation defined as the feature circle radius (FCR) is computed who has maximum when the radius of the region is equal to that of the circle. Synthetic images and natural images are used to test the capability of the proposed method. The experimental results indicate that the presented algorithm has excellent performance for detection of single circle, multiple circles, concentric circles and partial circles, and also good accuracy despite the presence of different noises.
Acta Automatica Sinica | 2011
Hongmin Liu; Zhiheng Wang; Chao Deng; Li-Qin Jia
Feature detection is one of the classic topics in the field of image processing;however,only little research has been made on polygon detection problem.Focusing on this problem,this paper presents a simple and effective method for polygon detection,called polygon detection method based on meta-representation.The main idea of this method is as follows:firstly,the key-points and their local edge directions are detected,and point metas (1-D metas) can be defined using the position and direction information of the key-points;then,line metas (2-D metas) can be attained by combining any two point metas that satisfy the constraint conditions;thirdly,a line meta and a point meta which satisfy the constraint conditions can be combined into a 3-D meta or triangle,and thus triangle detection is achieved.Similarly,considering an n-D meta (n ≥ 2) and a point meta satisfying the constraint conditions,(n + 1)-D meta or n + 1 sided polygon can be constructed,and thus polygon detection is achieved.Experiments show that the polygon detection method based on meta-representation can perform effectively and accurately for polygon detection.Besides,the meta-representation method proposed in this paper can provide an idea for detecting other graphics consist of line segments.
international conference on wavelet analysis and pattern recognition | 2010
Chao Deng; Hongmin Liu; Zhiheng Wang
A new noise removal algorithm based on improved neural network, is applied to remove the impulse noise of the digital images. First of all, an improved neural network is used to detect the noise-pixels and distinguish it from noise-free pixels efficiently; Second, the noise-pixels are replaced further by the suitable pixel which has the most local similarity; Finally, the output is the combination of the noise-free pixels and the suitable pixel. The proposed algorithm is capable of removing the impulse noise effectively. At the same time it can keep more image details well. Experiential results show that the new algorithm is more improved than the conventional filters.
international conference on machine learning and cybernetics | 2010
Hongmin Liu; Zhiheng Wang; Chao Deng
Feature matching plays an important role in many applications, including 3D reconstruction, object recognition and video understanding. Point matching has made great progress recently, while it has made little progress in the fields of line and curve matching. By computing statistics of point descriptors constructed at each edge points, this paper develops a novel method for extending point descriptors to construct descriptors for lines, curves and regions. Experiment results exhibit that our method can perform good and robust on real images.
international conference on wavelet analysis and pattern recognition | 2011
Hongmin Liu; Zhiheng Wang; Chao Deng; Li-Qin Jia
This paper presents a method for the square detection based on distance distribution of the edge points of the image. The orientation line of the edge point is defined first, then for each pixel of the image, the distances between the pixel and the orientation lines of the edge points in the defined neighborhood of the pixel are computed to obtain the feature length and feature energy of the pixel, and further attain the feature length distribution map and feature energy distribution map of the whole image. Local maxima detection in the feature energy map is subsequently performed to find the potential center of the squares, and the corresponding feature length of the found potential center is also obtained. Additive steps are performed to verify and determine the centers of the squares. Synthetic images and natural images are used to prove the capability of the presented method for square detection. The experiments exhibit good performance for detecting the single square or the nested ones.
CSISE (2) | 2011
Chao Deng; Xiao-rui Liu; Hongmin Liu; Zhiheng Wang
In view of the problem of a priori SNR estimation in speech enhancement, a novel algorithm is proposed in this paper. Through adding momentum term, the presented algorithm maintains the advantages of indirect decision algorithm in effectiveness, and improves the tracking speed of instantaneous SNR with the elimination of musical noise. Simulation results demonstrate that the algorithm possesses good performance in many kinds of noise conditions.
Archive | 2010
Li-Qin Jia; Zongpu Jia; Hongmin Liu; Zhiheng Wang; Xiao Xue
Archive | 2011
Hongmin Liu; Zhiheng Wang; Zongpu Jia
Archive | 2012
Hongmin Liu; Zhiheng Wang; Zongpu Jia