With the development of computer technology, digital imageprocessing technologies have been applied to many areas of real life, blurred image restoration also has a rapid development, which gives a certain basis and ...
详细信息
With the development of computer technology, digital imageprocessing technologies have been applied to many areas of real life, blurred image restoration also has a rapid development, which gives a certain basis and conditions in the bridge crack detection. Aiming at the blurred crack image generated by camera shake, the paper studies the motion blur image restoration algorithms, and explores the parameter estimation methods of motion blur, where the direction and scale of the blur kernel function are estimated from the spectrum of the blurred image. The paper uses different image quality evaluation standards to compare the output, which can choose the best results and gets a more accurate point spread function. This method can obtain clearer crack images and provide more accurate crack information for bridge project research.
image segmentation is a vital task in imageprocessing/computer vision. However, no universally accepted quality measure exists for evaluating the performance of various segmentation algorithms or even different param...
详细信息
image segmentation is a vital task in imageprocessing/computer vision. However, no universally accepted quality measure exists for evaluating the performance of various segmentation algorithms or even different parameterizations of the same algorithm. This paper proposes a new segmentation evaluation measure, based on the fusion of HOG and Harris features, thus we call it the H2. It exploits local shape, corner and edge information to evaluate the similarity between a given segmentation and its respective ground truth, and thus belongs to the category of supervised evaluation measures. The results obtained from our experiments show accuracy of up to 95% for the H2.
Pre-processing steps are critical in automated image analysis systems developed to aid in diagnosis of skin lesion images. The main areas of concern include, but are not limited to, hair on the skin, variations in ill...
详细信息
Pre-processing steps are critical in automated image analysis systems developed to aid in diagnosis of skin lesion images. The main areas of concern include, but are not limited to, hair on the skin, variations in illumination and skin tone, and alignment of successive skin images. These artifacts can partially or completely obstruct a lesion being analyzed causing errors in classification or diagnosis. This paper focuses on an independent quantitative evaluation of an open source hair removal algorithm [1]. The different input parameters to the algorithm were tested to determine their optimal values. Percent error and signal to noise ratio are utilized as the error metrics for the experimental results. Other essential pre-processing steps are considered and provided at the end of this paper.
Space field has experienced vigorous advancement with respect to evolution of vision system, image storage and processing. Real time imageprocessing has become one of the most important tools for navigation and landi...
详细信息
ISBN:
(纸本)9781538630051
Space field has experienced vigorous advancement with respect to evolution of vision system, image storage and processing. Real time imageprocessing has become one of the most important tools for navigation and landing for planetary and lunar missions. Information of horizontal velocity with high accuracy will be required to do accurate pin point landing. For the testability of such a system as well as to have the understanding of Lander dynamics, prior landing image sequences are required to initially testing the algorithm [1]. This paper deals with FPGA based processing on image sequence to find the relative velocity and also implements image Storage in a microSD card. Landing sequence is stored in SD card and post landing they are downloaded. These images provide a very useful information about lighting, lander dynamics to fine tune algorithm for future mission. Besides this image data serves the testability during various phases of testing during development.
In this paper is introduced a facial recognition system, based on mathematical methods, developed to act as the presence of students identifier in a classroom. It uses a recognition process in which seeks to extract r...
详细信息
In this paper is introduced a facial recognition system, based on mathematical methods, developed to act as the presence of students identifier in a classroom. It uses a recognition process in which seeks to extract relevant information from an image, to encode and compare them with another images of faces stored in an image database. Such image information representing a set of characteristics showing the variations between images of the obtained faces and contained in the image database. The Facial Recognition System consists of two processing modules: a training phase and a test phase, applied to a group of students in order to verify the utility of this algorithms for people recognizing.
This paper deals with real time segmentation of traffic images using a Mask R-CNN model. The aim is to improve the performance of real time image segmentation, so that it can be effective even with noisy images captur...
详细信息
This paper deals with real time segmentation of traffic images using a Mask R-CNN model. The aim is to improve the performance of real time image segmentation, so that it can be effective even with noisy images captured by traffic cameras. The approach developed here comprises of image preprocessing, object detection followed by segmentation. Mask R-CNN model not only segments the image, but also surrounds the image with bounding boxes and assigns class names to the individual objects e.g. Car, Truck, Bus, Bicycle, Person etc. The model is trained with the annotated MS COCO Training dataset. To improve the performance of Mask R-CNN over noisy images, here pre-processingalgorithms like Non Local Means (NLM) filter denoising and Median filter denoising are used. The testing is carried out on a subset of MS COCO Test dataset which comprises of only traffic images. The improved performance is demonstrated using parameters: increased correct object detections and corresponding confidence value, reduced incorrect object detections and corresponding confidence value, and an overall enhanced segment mask area accuracy.
This paper presents parallelization strategies for the implementation of imaging algorithms for synthetic aperture radar (SAR). Great emphasis is placed on time-domain based algorithms, namely the Global Backprojectio...
详细信息
ISBN:
(纸本)9781538631706
This paper presents parallelization strategies for the implementation of imaging algorithms for synthetic aperture radar (SAR). Great emphasis is placed on time-domain based algorithms, namely the Global Backprojection Algorithm (GBP) and its accelerated version, the Fast Factorized Backprojection Algorithm (FFBP). Multi-core platforms are selected for implementation as some combine good performance results with moderate power consumption. The implemented algorithms support several types of parallelization, as the stages of the algorithms can be handled sequentially or interleaved. For the GBP algorithm three different data distribution schemes are investigated. For the FFBP algorithm a successive stage calculation method is compared with a combined calculation method. The performance is exemplary evaluated on the low cost/energy, yet powerful multi-core platform Odroid-XU4. All parallelization strategies show an almost linear speed-up with the number of used cores. Even though a specific multi-core platform is investigated, the design decisions are applicable for general multi-core architectures.
Video sensors constitute a great innovation in the automotive sector and road safety as they contribute to the development of driver assistance systems. These video systems use imageprocessing techniques to inform dr...
详细信息
ISBN:
(纸本)9781509058150
Video sensors constitute a great innovation in the automotive sector and road safety as they contribute to the development of driver assistance systems. These video systems use imageprocessing techniques to inform drivers of impending dangers. One such development is the Lane Departure Warning System (LDWS) which play a key role in the prevention from accidents. The main function of this system is the detection of lane boundary lines using artificial vision. In this paper, we present a feature-based method for lane detection. We simplify the process of edge detection by using a horizontal differencing filter. The detected edge points are grouped into lines with a modified Hough transform.
imageprocessing and analysis is a useful tool for monitoring of activated sludge wastewater treatment plants. However, its effectiveness is dependent on performance of the segmentation algorithms. The activated sludg...
详细信息
ISBN:
(纸本)9781509011810
imageprocessing and analysis is a useful tool for monitoring of activated sludge wastewater treatment plants. However, its effectiveness is dependent on performance of the segmentation algorithms. The activated sludge wastewater plant can be monitored by imageprocessing and analysis of images acquired through microscope using bright field microscopy and phase contrast microscopy. In this paper, we have investigated three segmentation algorithms which are channel based segmentation, edge based segmentation and Bradley based segmentation. The performance of the algorithms is assessed using the performance metric of accuracy. Forty gold approximations of ground truth images are manually prepared for comparing with the result for segmentation. Half of the forty images are acquired at lOx and rest at 20x objective magnification of the microscope. Edge based segmentation gives better results compared to other algorithms with accuracy of 0.972.
In multiview systems, color plus depth format builds 3D representations of scenes within which the users can freely navigate by changing their viewpoints. In this paper we present a framework for view synthesis when t...
详细信息
In multiview systems, color plus depth format builds 3D representations of scenes within which the users can freely navigate by changing their viewpoints. In this paper we present a framework for view synthesis when the user requests an arbitrary viewpoint that is closer to the 3D scene than the reference image. On the target image plane, the requested view obtained via depth-image-based-rendering (DIBR) is irregularly structured and has missing information due to the expansion of objects. We propose a novel framework that adopts a graph-based representation of the target view in order to interpolate the missing image pixels under sparsity priors. More specifically, we impose that the target image is reconstructed with a few atoms of a graph-based dictionary. Experimental results show that the reconstructed views have better PSNR and MSSIM quality than the ones generated within the same framework with analytical dictionaries, and are comparable to the ones reconstructed with TV regularization and linear interpolation on graphs. Visual results, however, show that our method better preserves the details and results in fewer disturbing artifacts than the other interpolation methods.
暂无评论