Advanced Materials Research Vols. 760-762

Abstract: Wavelet transformation was introduced into the image enhancement step of infrared thermal wave NDT (nondestructive testing). This paper first explained the theory of image enhancement of wavelet transformation and then devised the enhancing process, where to get the best enhancing effect, this paper studied the selection of the threshold value and enhancement factor of the enhancement function to draw the conclusion that threshold value T mainly has impact on noise reduction, enhancement factor k₁ on the enhancing effect of transition area, and enhancement factor k₂ on image contrast, and also described the parameter selection method; on this basis, this paper finally selected threshold value T 600, k₁ 10, and k₂ 0.15 to carry out enhancement processing on the infrared thermogram of a composite and by comparison with the original image proved the effectiveness of this method.

Abstract: Removed at authors request

Abstract: Facial recognition is an important research topic in biometrics and has wide applications in pattern recognition and computer vision. This paper aims at providing interested readers with recent progresses in human face detection, tracking, video-based face recognition, and 3D+2D hybrid face recognition. For this purpose, it mainly focuses on those state-of-the-arts methods and technologies that emerged in recent previous few years. Most existing methods in this area are still-image based which do not utilize motion information; whereas most video-based methods work only in 2D video sequences, which are subject to pose and illumination variations. The recent emergence of 3D video cameras capable of producing range image sequences and corresponding texture image sequences simultaneously allows for the possibility of facial recognition in a 3D+2D video-based scenario. In view of this fact, a scheme of face detection, tracking and recognition process in 3D video-based manner is also proposed in this paper with further concerns addressed.

Abstract: Data hiding is one of the most important ways to achieve secure data transfer. Most spatial data hiding methods will make the host image distortion and the embedded data can be detected by stenography analysis. A novel data hiding method is proposed in this paper. This method first disturbs the raw critical data, and then embedded them only into areas of color changes much in the host image. Experiments show that this method can reduce the distortion of the host image and decrease the appreciability of the embedded data.

Abstract: Image segmentation in medical image processing has been extensively used which has also been applied in different fields of medicine to assist doctors to make the correct judgment and grasp the patient's condition. However, nowadays there are no image threshold segmentation techniques that can be applied to all of the medical images; so it has became a challenging problem. In this paper, it applies a method of identifying edge of the tissues and organs to recognize its contour, and then selects a number of seed points on the contour range to locate the cancer area by region growing. And finally, the result has demonstrated that this method can mostly locate the cancer area accurately.

Abstract: Indoor scene understanding based on the depth image data is a cutting-edge issue in the field of three-dimensional computer vision. Taking the layout characteristics of the indoor scenes and more plane features in these scenes into account, this paper presents a depth image segmentation method based on Gauss Mixture Model clustering. First, transform the Kinect depth image data into point cloud which is in the form of discrete three-dimensional point data, and denoise and down-sample the point cloud data; second, calculate the point normal of all points in the entire point cloud, then cluster the entire normal using Gaussian Mixture Model, and finally implement the entire point clouds segmentation by RANSAC algorithm. Experimental results show that the divided regions have obvious boundaries and segmentation quality is above normal, and lay a good foundation for object recognition.

Abstract: In distance education network transmission process, because transmission distance is too long, transmission network will be affected by complicated external environment factors, which leads to network failure and failure in remote education video image formation, and finally causes unsmooth transmission. This paper puts forward a distributed network fault detection technology to perform fault detection for remote education transmission network nodes and characteristic analysis of the use of network fault by using genetic neural network, accurately locate fault node area so as to realize the remote education networks fault detection. Experiments show that this method can avoid distance education network fault resulted from long transmission distance and improve the transmission efficiency of remote education video image.

Abstract: Using remote sensing technique to determine coastline's position has been received vital attention. This paper presents a novel approach for detecting coastline of remote sensing image based on K-Means cluster and Distance Transform algorithm. K-Means cluster algorithm divides the image into two regions-water and land area. Then to extract the sea area by distance transfoming. Finally, the coastline will be detected by edge traking. Results showed that the method proposed in this paper have good performance in accuracy and completeness.

Abstract: This paper proposes extracting black line with small-scale method in the self-tracing car system based on OV6620 as the video sensor unit. Describe the hardware design of OV6620 and other two kinds of image signal extracting methods, which are to extract black line from left to right, and extract black line from middle to both sides. The testing shows that extracting black line with small-scale method has good stability, accuracy, and anti-interference.

Abstract: ris location is one of the key steps of iris recognition system. Non-ideal iris image has some problems, such as eyelid and eyelash occlusion, low contrast of iris and sclera, uneven illumination, and so on. Because of that, its difficult to identify the boundary, especially the exterior boundary. Therefore, this paper proposes a method based on the improved Hough Transform. First, use the minimum method to find the datum point in the pupil, after that identify the valid area of the interior boundary base on that point. Apply the improved Hough Transform to that valid area to identify the interior boundary of the iris image. Then regard the center of the interior circle as our new datum point, use the same method to identify the exterior boundary. Experiment results show that our algorithm has higher accuracy than traditional method on the non-ideal iris image segmentation.