Applied Mechanics and Materials Vols. 333-335

Abstract: Image segmentation techniques currently used for X-ray inspection in pharmaceutical industry suffer from some limitations. The object in an image is close to the background and its contours are weak or blurred because of the X-ray imaging characteristic. Based on our research of X-ray inspection, a simple and efficient image segmentation method is proposed in this paper. It is implemented by treating the image and desired contours as three dimensional surface and holes respectively in order to simplify the model of segmentation, and making use of surface fitting and image subtraction to extract the target region efficiently. The novelty of this approach is that we need less selection of parameters to extract contours with low contrast by surface fitting. Experiments on real X-ray images demonstrate the advantages of the proposed method over active contour model (ACM) and Chan_Vese model (CV model) in terms of both accuracy and efficiency on fixed condition.

Abstract: An algorithm of spatio-temporal combining video denoising based on structural similarity is proposed for the video surveillance system. By motion detection to multi-frame images with block structural similarity, this algorithm can adaptively distinguish the still regions and motion regions of video image. Temporal weighted average filter to the still regions and spatial ANL filter to the motion regions are used separately. Experimental results show that the proposed algorithm can improve the image quality and greatly reuduce the computation time.

Abstract: In this paper, we present a dynamic gesture recognition system. We focus on the visual sensory information to recognize human activity in form of hand movements from a small, predefined vocabulary. A fast and effective method is presented for hand detection and tracking at first for the trajectory extraction. A novel trajectory correction method is applied for simply but effectively trajectory correction. Gesture recognition is achieved by means of a matching technique by determining the distance between the unknown input direction code sequence and a set of previously defined templates. A dynamic time warping (DTW) algorithm is used to perform the time alignment and normalization by computing a temporal transformation allowing the two signals to be matched. Experiment results show our proposed gesture recognition system achieve well result in real time.

Abstract: A digital character recognition method is presented based on BP Neural Network. This paper preprocesses the digital character image and extracts character feature, then uses BP Neural Network to recognize digital character. Back Propagation algorithm seeks network weights to minimize training error in the solution space. A network with hidden layer is created at first, then an input sample vector is sent to network input terminal and the square error E between output values and training sample object output values is calculated. Above process is repeated for input samples of training sets until the error is reduced within the limits of the threshold. The results show that the method presented has good accuracy, quick speed and strong robustness for realtime application.

Abstract: Vacuum circuit breaker is one of the major critical components that controls and protects the power system. The breaking capability of the vacuum switch directly affects the safety of the power system. However, the breaking capability of the vacuum switch is determined by the morphological characteristics of the vacuum arc images. In this paper, in order to research the variation of the vacuum arc, a high-speed CMOS camera system is used to collect the vacuum arc images within a short gap (7mm), the morphological characteristics of the vacuum arc images are extracted based on the digital image processing. At last, the variation of the arc is analyzed.

Abstract: In order to adapt to the requirements of intelligent video monitoring system, this paper presents an ARM-Linux based video monitoring system for face detection. In this system, an ARM processor with a Linux operating system was used, and the USB camera was used to capture data, and then the face detection was conducted in the ARM device. The OpenCV library was transplanted to Linux embedded system. The algorithm of face detection was realized by calling the OpenCV library. Specially, adaboost algorithm was chose as the face detection algorithm. Experimental results show that the face detection effect of the system is satisfactory and can meet the real time requirement of video surveillance.

Abstract: Image compression for image storage and transmission are very necessary. In order tomore save memory space and more be used as a compressed representation of a image, a image ismapped into a graph, we can use the labeling of graph to compress and give a new graphs compressedrepresentation.

Abstract: Nonlocal regularization for image restoration is extensively studied in recent years. However, minimizing a nonlocal regularization problem is far more difficult than a local one and still challenging. In this paper, a novel nonlocal total variation based algorithm for image deblurring is presented. The core idea of this algorithm is to consider the latent image as the fixed point of the nonlocal total variation regularization functional. And a split Bregman method is proposed to solve the minimization problem in each fixed point iteration efficiently. By alternatively reconstructing a sharp image and updating the nonlocal gradient weights, the recovered image becomes more and more sharp. Experimental results on the benchmark problems are presented to show the efficiency and effectiveness of our algorithm.

Abstract: t is quite constrained for us to use some other input devices to communicate with computers. In this paper, we integrate human-computer interaction technologies with handwritten Chinese character recognition strategies using depth image information provided by Kinect sensor to realize an unconstrained handwritten character recognition system, which only uses our hand as input device. We predefine several hand gestures as instructions, and for the recognition of these hand gestures, we calculate the contour and fingertips of the hand used for writing using depth image taken by Kinect. By mimicking the functionalities of the computer mouse only using our hands, we can write freely in the air and get the original character image. After Gaussian blurring and normalization, we adopt some classic handwritten character recognition schemes to accomplish the recognition task. Experiments show that the system gives a good result.

Abstract: To deal with locally narrow, low-contrast and spatially varying intensity in segmentation of the computed tomography angiographic (CTA) data, a deformable model with newly proposed edge measures was presented for segmenting coronary arteries. The edge measures were derived from the refined vesselness measures of brightness and multi-scale filtering responses, i.e., vesselness and scale. The initial vessel region and boundary region was derived from the multi-scale filtering responses, from which the statistical information of vessel appearance was attained to yield brightness measure. As compared with the multi-scale filtering responses, the refined vesselness measures could effectively suppress non-vascular background while preserving vessel-like structure. Finally, the new edge measures were embedded into deformable model, resulting in better artery segmentation.