Applied Mechanics and Materials Vols. 536-537

Abstract: Visual image, as a kind of rich content and performance of multimedia information, has been tremendously popular for a long time. Using text-based Image Retrieval TBIR (Text-based Image Retrieval, TBIR) during retrieval will provide keywords and description of the image text matching, operation simple and quick. The defects of TBIR also, however, there are the following: (1) image library image all the need for manual annotation, time-consuming and laborious with subjective factors; (2) image semantics is rich, simple key words cannot fully express its meaning and accurate. Image Retrieval Based on regional RBIR (Region-based Image Retrieval, RBIR) first of all, by using image segmentation method, divides an image into several different regions. At last image matching is converted to match between the regions. We just need to user submits a retrieval image, greatly reducing the user's retrieval burden.

Abstract: Object tracking is one of the most important components in numerous applications of computer vision. Much progress has been made in recent years. The Tracking-Learning-Detection (TLD) algorithm achieves excellent performance on a set of challenging video sequences. However, classical TLD algorithm fails to track non-rigid pedestrian as complicated appearance, varying viewpoints, shape changes and occlusions. In this paper, we follow the TLD and propose a robust KMD framework which consists of the Kalman filter tracker, random ferns pedestrian detector and the pedestrian model. During the tracking process the tracker and detector are complementary: the Kalman filter tracker predicts the motion, the pedestrian detector searches the best-match appearance and the pedestrian model combines performance of both to determine final result and generate new samples for learning. Experimental results show our framework comparatively improves performance for pedestrian tracking in surveillance videos.

Abstract: This paper takes the following measures to solve the problem of 3D reconstruction. Camera calibration is based on chessboard, taking several different attitude images. Use corner point coordinates by corner detection to process camera calibration. The calibration result is important to be used to correct the distorted image. Next, the left and right images should be matched to find out the object surface points’ imaging position respectively so that the object depth can be calculated by triangulation. According to the inverse process of projection mapping, we can project the object depth and disparity information into 3D space. As a result, we can obtain dense point cloud, which is ready for 3D reconstruction.

Abstract: Segmentation of brain MRI in an image sequence is one of the most challenging problems in image processing, while at the same time one that finds numerous applications. In this paper, we propose a robust multi-layer background subtraction technique and seed region growing approach which takes advantages of local texture features represented by local binary patterns (LBP) and photometric invariant color measurements in RGB color space. Due to the use of hybridization of layer-based strategy and seed region growing approach, the approach can model moving background pixels with quasiperiodic flickering as well as background scenes which may vary over time due to the addition and removal of long-time stationary objects. The experiment results prove that in the view of the brain MRI segmentation, this algorithm provides fast segmentation with high perceptual segmentation quality.

Abstract: In the acquisition process of palmprint image, the image is often subject to interference from the outside noise.This noise will affect the future of Palmprint feature extraction. Denoising of wavelet is a common method for denoising of palmprint image. The defect of traditional soft and hard threshold functions in image denoising is analyzed. An improved threshold function is proposed on the basis of soft and hard threshold functions. The conventional hard, soft threshold functions and the improved function are used respectively in image denosing. Experiments prove that the improved wavelet threshold function is better than conventional soft, hard threshold function in denoising effect.

Abstract: This paper analyzed the reconstruction of cross-cut shredded paper coming from a same printed page, built two association degree optimization models (including edge association degree optimization model and the line space one) based on the characteristics of character information on the edge and line space, and provided the concrete solving process and algorithm for the reconstruction of shredded paper. Moreover, the solving process and algorithm were applied in a reconstruction experiment for pieces of shredded paper from China Undergraduate Mathematical Contest in Modeling (CUMCM 2013B). The results showed that this reconstruction method was effective.

Abstract: In the research of identity recognition based on lip motion features, there are limitations for the existing algorithms of lip characteristic parameters extraction. This paper uses the strategy of lip static/dynamic geometric features fusion, designs the lip feature parameter extraction program based on interpolation, and implements the major aspects of processing algorithm of the program. The solution is based on the speaker's key six primitives spelling lip sequence image, firstly generates the lip key point coordinates in the image, then based on Lagrange interpolation obtains function curve coefficient of upper and lower lips' key points , lastly the two curve coefficients are combined to form lip motion feature information of human speaker's some specific sounds; Simulation results show that the extraction of characteristic parameters of the program not only have a high efficiency and availability, but also have the advantages of good storage.

Abstract: Based on Bandelet transform and Multi-fractal, the metal fracture surface image denoising is described in the paper. Using multi-fractal analysis, eight multi-fractal spectrum values which are used as shape characteristic parameter of metal fracture surface are extracted. Experimental results showed that different metal fracture surfaces have different shape characteristics and the same kind of metal fracture surfaces have similar shape characteristics. So the shape characteristic parameter of metal fracture surface can be used to study the fracture surface recognition and as the sample of failure analysis.

Abstract: By using Hamilton improved circle, transversely and longitudinally cut up paper join together. By scanning image and image extraction techniques, the images of shredded paper are obtained. The process of shredding splicing is mainly divided into two parts, the image preprocessing and edge matching. Then the images conver into grey value matrixes and binarization are implemented on the image with appropriate threshold. The images of shredded paper are classified into nineteen lines. With the shortest Euclidean distance between the edge of shredding and Hamilton improved circle, shredded paper is transversely spliced. The test shows that accuracy is as high as 100%.

Abstract: This paper introduces the IP network video monitoring system for monitoring and management of industrial field. All the signals in this monitoring system are transmitted through the network, without additional wiring. It not only can transmit images, alarm information, also the focus of a number of functions, such as remote monitoring, video playback, query system settings and user management. The system overall design principle, structural design, functional requirements, the main function module design are described. Through the monitoring system, managers can whenever and wherever real-time monitoring equipment working state, remote access management, provide a full range of decision support for production.