Qieshi Zhang

Face detection and tracking in color images using color centroids segmentation

Human face detection plays an important role in many application areas such as video surveillance, human computer interface, face recognition, face search and face image database management etc. In human face detection applications, face region usually form an inconsequential part of images. Consequently, preliminary segmentation of images into regions that contain “non-face” objects and regions that may contain “face” candidates can greatly accelerate the process of human face detection. Color information based methods take a great attention, because colors have obviously character and robust visual cue for detection. This paper proposed a new method based on RGB color centroids segmentation (CCS) for face detection. This paper include two parts, first part is color image thresholding based on CCS and the second part is face detection based on region growing and facial features structure character combined method. The experimental results show the ideal thresholding result and better than the result of other color space analysis based thresholding methods. Proposed method can conquer the influence of different background conditions, position, scale instance and orientation in images from several photo collections and database; the effect is also better than existing skin color segmentation based methods.

Automatic road sign detection method based on Color Barycenters Hexagon model

Road sign detection is one of the major concerned topics in the field of driving safety and intelligent vehicle. In this paper, a novel model based on Color Barycenters Hexagon (CBH) is proposed and used to detect road sign usefully. In CBH model, full color images are calculated the color barycenters and get the barycenters region, then automatic select the idea threshold curves to separate the region of interest (ROI) of barycenters aiming to detect the road sign. Because of the practically images have many noise, and the existing color space cannot separate the ROI ideally. The proposed CBH model can thresholding the principal color of ROI and have high robust. With suitably thresholding and operations, road sign on various scene images can be detected.

Adaptive Histogram Analysis for Image Enhancement

One image processing application is to reconstruct the original scene from the low quality images. Considering the idea histogram distribution can reflect good vision effect. So many histogram analyzing based methods have been studied recently. However, some methods require users to set some parameters or condition, and cannot get the optimal results automatically. To overcome those short come, this paper presents an Adaptive Histogram Separation and Mapping (AHSM) method for Backlight image enhancement. First, we separate the histogram by binary tree structure with the proposed Adaptive Histogram Separation Unit (AHSU). And then mapping the Low Dynamic Range (LDR) histogram partition into High Dynamic Range (HDR). By doing this, the excessive or scarcity enhancement can be avoid. The experimental results show that the proposed method can gives better enhancement results, also compared with some histogram analyzing based methods and get better results.

Improved Optical Model Based on Region Segmentation for Single Image Haze Removal

In this paper, we propose a novel method to recovery the haze-free image based on the improved optical model with the single image based depth estimation. In this work, the objective transmission and distance transmission are present to improve the optical model for obtaining the haze-free image. Firstly, the color clustering method is used to segment the image into several regions by color similarity. Then, the graph-based segmentation is used to calculate the depth information. Next, the atmosphere light is estimated according to the distance transmission. And finally, the improved optical model is used to estimate the haze-free image. The experimental results show that our method is more effective and able to get better results than other single image based methods.

Disparity refinement with stability-based tree for stereo matching

This paper proposes a disparity refinement method with stability-based tree. By developing stability-based tree to evaluate and reconstruct support regions for error parts, the proposed method achieves effective performance in removing outliers. This approach further improves the quality of raw disparity map in stereo matching, which makes the local methods results comparable to the global ones. Experiments exhibit that the proposed method reduces more than 70% aggregation time compared with traditional tree method without loss of accuracy. It also outperforms existing disparity refinement methods in removing large error parts.

Robust road lane detection using extremal-region enhancement

Road lane detection is a key problem in advanced driver-assistance systems (ADAS). For solving this problem, vision-based detection methods are widely used and are generally focused on edge information. However, only using edge information leads to miss detection and error detection in various road conditions. In this paper, we propose a neighbor-based image conversion method, called extremal-region enhancement. The proposed method enhances the white lines in intensity, hence it is robust to shadows and illuminance changes. Both edge and shape information of white lines are extracted as lane features in the method. In addition, we implement a robust road lane detection algorithm using the extracted features and improve the correctness through probability tracking. The experimental result shows an average detection rate increase of 13.2% over existing works.

Color image segmentation based on wavelet transformation and S OFM neural network

Image segmentation, which is the first essential and fundamental issue in the image analysis and pattern recognition, is a classical difficult problem in the image processing. The color images, which possess more visual information than the gray images do, have aroused more and more attentions. In the medical imaging system, according to the different absorbency of different tissues, the staining method is often used to get the color image which provides more abundant information for diagnosis. As for the automatic analysis system of kidney-tissue image stained by Periodic Acid Schiff (PAS), the correct segmentation of glomerulus is an important step. A layer- color clustering segmentation method based on wavelet transformation and self-organizing feature map neural network (SOFM) is proposed in this paper. Firstly, the wavelet transformation is applied to the original images to get the low frequency images to improve the running efficiency. Secondly, the disordered method based on random number is performed to improve the performance of SOFM. Thirdly, the layer-color clustering using SOFM is executed until the final error can meet the need of the average color error (ACE) and then the clustered image and the palette can be acquired. Finally, based on the histogram of palette, the glomerulus can be segmented from the kidney-tissue image correctly. Experimental results show the good performance of this method.

Two-stage cross-based stereo disparity refinement

This paper proposed a disparity refinement method based on two-stage cross. First stage is anti-texture cross-based support region construction to build proper support regions for error pixels without being influenced by texture. Based on the support regions, second stage of the method is proposed, which is called weighted cross-based updating method. The experiments show that the proposed method could build the support region accurately and improve the accuracy of the disparity map in final results with fast speed, compared to other tree-based algorithms. It also outperforms the existing disparity refinement methods in preserving the boundaries of objects in the final disparity map.

Adaptive sampling and wavelet tree based compressive sensing for MRI reconstruction

Magnetic Resonance Imaging (MRI) has been widely used in medical diagnose because of its non-invasive manner and excellent depiction of soft-tissue changes. Recently, the compressive sensing (CS) theory has been applied to reconstruct the MR image from highly down-sampled k-space data, which can reduce the scanning duration. To obtain useful information as much as possible with the same sampling rate, a weighted sampling strategy is studied. Moreover, based on the advantage of CS, a Wavelet tree based reconstruction approach is proposed. The experimental results demonstrate that the proposed method is preferable to other methods.

A novel color space based on RGB color barycenter

Color space is one of the bases in the image processing area. Suitable color space can give the suitable description of colors for variant processing. However, in the image processing area, the existing color space cannot show the suitable distribution in color and lightness. In this paper, a novel color space based on RGB color barycenter (RGB-CB) is proposed to describe the color and lightness more intuitively. To prove the effectiveness of the proposed color space, YUV, HSV, L*a*b*, and IPT color spaces are discussed and compared. Experimental results show the proposed color space can perform better effect than other color space in image processing.