Liu Wanyu

A camera calibration method based on neural network optimized by genetic algorithm

The calibration of camera is to determine the relation between the two dimensional (2D) image coordinates and the corresponding three dimensional (3D) world points, and is the basis of vision inspection system. This paper presents a new neurocalibration approach based on the neural network optimized by Genetic algorithm (GA) for camera calibration. Unlike other existing approaches based on neural network, our calibrating method can give a theoretical optimization solution for the problems in using neural network. We use GA to optimize the structure, the connection weights and the threshold values of the neurons of the neural network. Though the training time of our method is longer than the BP neural network, the experiments results show that the method we proposed is feasible, robust and effective.

The research of video tracking based on improved SIFT algorithm

In order to effectively solve the problem on the moving object tracking in the complex environment, a method that adopts the Scale Invariant Feature Transform (SIFT) is presented in this paper. Firstly, a searching window of narrow size is set. Secondly, extreme points are detected in scale space and the main direction of the feature points is calculated. Thirdly, in order to find the matched keypoint descriptors, we set a threshold, if the ratios of the nearest distance to the second nearest distance in current frame are below the threshold, the keypoints will be kept. It is a better solution to the shortcoming of rather time-consuming for traditional SIFT algorithm in determining descriptions of the large number of keypoints. The experimental results verified the SIFT superiority properties of the scale-invariant, rotation-invariant and other image transformations, and the good tracking results are demonstrated for the moving car in spite of the background is changing as well.

A Method of Pulmonary Nodules Detection with Support Vector Machines

Lung cancer is one of the deadly and most common diseases in the world. Many methods have been proposed to avoid radiologists fail to diagnose small pulmonary nodules. Recently, support vector machines (SVMs) had received an increasing attention for pattern recognition. We present a computerized system aimed at pulmonary nodules detection; it identifies the lung field, extracts a set of candidate regions with a high sensitivity ratio and then classifies candidates by the use of SVMs. We performed several experiments with different kernels and differently balanced training sets. The results obtained show that cost-sensitive SVMs trained with unbalanced data sets achieve promising results in terms of sensitivity and specificity. The studies have shown a high potential for implementation of this system in clinical practice as a computer aided diagnosis (CAD) tool.

A robust algorithm based on nonstationary degree for ultrasonic data enhancement

Compared with other medical imaging modalities, ultrasound imaging has its own advantages. The three-dimensional (3-D) ultrasound stereo visualization technique has a broad promising future for its ability superior to traditional two-dimensional (2-D) ultrasound image, and it helps to understand the complex structures of tissues as well as to measure tissular volumes. However, it is often difficult to interpret the 3-D structure from acoustic data because of the speckle noise. To solve this problem, we propose a new enhancement algorithm which is based on calculating nonstationary degree of ultrasound data to improve the image quality. We observe data within a finite length window and then map them to an N-dimensional space where every point represents the observed data. According to the data features, we divide this space into two parts: the stationary subspace constituted by the stationary points, which represent a line in the space, and the nonstationary subspace is formed by the points out of the line. Then the nonstationary degree of a set of observed data is defined as the distance from the correspondent point to the stationary line. Thus, we can enhance the image since the nonstationary degree is larger on tissular border where features vary rapidly. Finally, the application of the proposed algorithm to real data of the liver of a rabbit is described. The results are shown by means of 3-D ultrasound stereo visualization, and the results demonstrate a significant improvement compared with the original image.

A Novel Algorithm for Ultrasonic Stereovision Data Enhancement

The three-dimensional (3-D) ultrasound stereo visualization technique has a broad promising future for its ability superior to traditional two-dimensional (2-D) ultrasound image, and it helps to understand the complex structures of tissues as well as to measure tissular volumes. However, it is often difficult to interpret the 3-D structure from acoustic data because of the speckle noise. To solve this problem, we propose a new enhancement algorithm which is based on calculating nonstationary degree of ultrasound data to improve the image quality. We observe data within a finite length window and then map them to an N-dimensional space where every point represents the observed data. According to the data features, we divide this space into two parts, and then the nonstationary degree of a set of observed data is defined as the distance from the correspondent point to the stationary line. Thus, we can enhance the image since the nonstationary degree is larger on tissular border where features vary rapidly. Finally, the application of the proposed algorithm to real data of a 12-week embryonic foetus is described. The results are shown by means of 3-D ultrasound stereo visualization, and the results demonstrate a significant improvement compared with the original image.