• Publications
  • Influence
A Modified FCM Algorithm for MRI Brain Image Segmentation
A modified FCM algorithm (called mFCM later) for MRI brain image segmentation is presented, realized by incorporating the spatial neighborhood information into the standardFCM algorithm and modifying the membership weighting of each cluster.
A novel image retrieval method based on hybrid information descriptors
A set of descriptors based on human visual perception mechanism designed to explore the internal correlations among different image feature spaces with image structure and multi-scale analysis are proposed.
A novel approach for segmentation of MRI brain images
A novel method for segmentation of brain tissues in MRI (magnetic resonance imaging) images is proposed in this paper. First, we reduce noise using a versatile wavelet-based filter. Subsequently,
An improved locality sensitive discriminant analysis approach for feature extraction
Improved Locality Sensitive Discriminant Analysis can not only preserve the local discriminant neighborhood structure of the data, but also pull the outlier samples more close to their class centers, which makes it outperform the original LSDA and some other state of the art algorithms.
A novel license plate localization method based on textural feature analysis
A four-stage search process is proposed in this paper: pre-processing, approximate region searching, region classification and skew correction using Radon transform, and results show that almost 96.1 % of input images are correctly localized on the average.
A Novel Moving Cast Shadow Detection of Vehicles in Traffic Scene
A novel approach which adequately considers color space information to detect moving cast shadows of vehicles in traffic videos is proposed, which achieves both high shadow detection and discrimination rates but takes on better performance than some state-of-the-art methods.
Maximum weight and minimum redundancy: A novel framework for feature subset selection
A novel filter framework is presented to select optimal feature subset based on a maximum weight and minimum redundancy (MWMR) criterion that can select the feature subset in which the features are most beneficial to the subsequent tasks while the redundancy among them is minimal.
Label propagation based semi-supervised non-negative matrix factorization for feature extraction
Extensive experimental results demonstrate that by propagating the label information and factorizing the matrix alternately, the proposed LpSNMF algorithm can obtain better performance than some other algorithms.
Plant Species Identification Based on Neural Network
A new method for plant species identification using leaf image that focuses on the stable features extraction of leaf, such as the geometrical features of shape and the texture features of venation, using SOM neural network to identify the plant species.
A novel image hiding approach based on correlation analysis for secure multimodal biometrics
Extensive experimental results demonstrate that the proposed approach not only provides good imperceptibility but also resists some common attacks and assures the effectiveness of network-based multimodal biometrics identification.