This paper presents three fully convolutional neural network architectures which perform change detection using a pair of coregistered images. Most notably, we propose two Siamese extensions of fully convolutional net...
详细信息
This paper presents three fully convolutional neural network architectures which perform change detection using a pair of coregistered images. Most notably, we propose two Siamese extensions of fully convolutional networks which use heuristics about the current problem to achieve the best results in our tests on two open change detection datasets, using both RGB and multispectral images. We show that our system is able to learn from scratch using annotated change detection images. Our architectures achieve better performance than previously proposed methods, while being at least 500 times faster than related systems. This work is a step towards efficient processing of data from large scale Earth observation systems such as Copernicus or Landsat.
In this paper, an electrical capacitance tomography (ECT) system for real-time measurement of solid contaminants in gas pipelines is presented. It consists of a ring of eight electrodes evenly distributed in the circu...
详细信息
In this paper, an electrical capacitance tomography (ECT) system for real-time measurement of solid contaminants in gas pipelines is presented. It consists of a ring of eight electrodes evenly distributed in the circular cross section of the probe. The speed-up enhancement is achieved using a field programmable gate array (FPGA) for the post-processing part of the system to accelerate the intensive matrix multiplications which are required in the image reconstruction algorithm. Experimental results on field-collected solid contaminants demonstrated the capability of the system to build in real-time two-dimensional cross-sectional images of the contaminants while giving an estimated measurement of their concentration. This allows the flow regime of the contaminants in the pipeline to be identified. Results also show that using Altera's Stratix v FPGA, 305 KLEs are required to achieve image reconstruction throughput up to 3233 frames/s for image size of 64 x 64 pixels. Simulation results were also conducted using finite-element method solver to assess the ECT probe for various image reconstruction algorithms (i.e., Linear back projection, Landweber, and modified Landweber algorithms). The results indicate a good matching with the experimental results.
The embedded multimedia devices which are designed to perform computationally intensive DSP applications such as imageprocessing and a large range of multimedia tasks. For improving the performance of image processin...
详细信息
ISBN:
(数字)9781538624401
ISBN:
(纸本)9781538624418
The embedded multimedia devices which are designed to perform computationally intensive DSP applications such as imageprocessing and a large range of multimedia tasks. For improving the performance of imageprocessingsystems, processingalgorithms should be implemented in hardware platforms. For real time applications, reconfigurable hardware in the form of Field Programmable Gate Arrays (FPGAs) are used which are able to provide high performance with low latency. The inherent reprogram ability of FPGAs gives them the flexibility of software while retaining the performance advantages of an application specific solution. In real time applications as image sizes grow larger, only real-time hardware systems have to be used with less software. This paper proposes a fully pipelined architecture implementation of skeletonization algorithm for 2-D gray scale images. 3x3 windowing operator is used for analyzing the pixel values. The proposed architecture is tested for image size of 8×8, but the approach discussed can be used for images of any size, as long as the FPGA memory will hold it. The implementation was carried out on Xilinx vertex 5 board.
Introduction: Biological systems which store medical images should have high reliability in real time. If the bit in an image flips or altered due to voltage ripples or wakes, the information content will be lost, and...
详细信息
Introduction: Biological systems which store medical images should have high reliability in real time. If the bit in an image flips or altered due to voltage ripples or wakes, the information content will be lost, and hence the medical diagnosis would be difficult for the clinicians. These bit errors may look like speckles in an image, and thus there is more demand for the despeckling algorithms. Methodology: In common, the median filter plays a crucial role to reduce the speckle noise or impulse noise. This study explores the implementation of the existing method and five novel proposed algorithms to mitigate bit errors in medical images. Here, the proposed techniques are mainly based on point processing, edge detection, and morphological processingalgorithms. The performance metrics like peak signal-to-noise ratio (PSNR), signalto- noise ratio (SNR), quality index (QI), degree of distortion (D), root-mean-square error (RMSE), structural similarity index (SSIM) and the execution time are used for the comparison of the existing method and the proposed algorithms. Results & Conclusion: The existing method and the proposed algorithms are implemented using Matlab R2016b. Simulation results show that the proposed algorithms provide better results than the current approach. For example, the fourth proposed algorithm provides better PSNR (47.28 dB) than others for the abdomen magnetic resonance imaging (MRI) scan image.
This paper presents the implementation of Particle filter based Object tracking using ReconOS on Reconfigurable Computing System. Multithreading can be used to improve the performance of complex imageprocessing algor...
详细信息
This paper presents the implementation of Particle filter based Object tracking using ReconOS on Reconfigurable Computing System. Multithreading can be used to improve the performance of complex imageprocessingalgorithms, but their sequential execution is a barrier which can be tackled by the effective use of FPGAs. In order to accomplish the desired performance, an operating system, ReconOS is used on an ARM based CPU-FPGA hybrid platform. ReconOS extends communication and synchronization primitives of operating systems like mutexes, semaphores, condition variable and message boxes to reconfigurable hardware. ReconOS provides the advantage of mapping the particle filter algorithm into reconfigurable hardware and accessing the data from software threads. Thus providing improved performance, portability and unified appearance as well as transparency to the object tracking application.
We propose a GPU-based approach to accelerate filtered backprojection (FBP)-type computed tomography (CT) algorithms by adaptively reconstructing only relevant regions of the object at full resolution. In industrial a...
详细信息
We propose a GPU-based approach to accelerate filtered backprojection (FBP)-type computed tomography (CT) algorithms by adaptively reconstructing only relevant regions of the object at full resolution. In industrial applications, the object's insensitivity to radiation as well as lack of inner motion allow for high-resolution scans. The large amounts of recorded data, however, pose serious challenges as the computational cost of CT reconstruction scales quartically with resolution. To ensure real-time reconstruction (i.e. faster processing than projection acquisition) for high-resolution scans, our method skips below-threshold voxels and monotonous regions inside the object. Our approach is able to speed up the reconstruction process by a factor of up to 13 while simultaneously reducing memory requirements by a factor of up to 71.
Increasing the effectiveness of training and training sessions is possible through the implementation of so-called biological feedback. Such feedback allows the teacher, or the instructor, to continuously monitor the ...
详细信息
Increasing the effectiveness of training and training sessions is possible through the implementation of so-called biological feedback. Such feedback allows the teacher, or the instructor, to continuously monitor the current psycho-emotional and functional state of the students. As a result, it becomes possible to adapt the style, pace, training mode and the volume of the material outlined, depending on the current receptivity and fatigue level of the listeners. The main element of systems that implement biological feedback in practice are remote non-contact technologies. Such technologies allow in a fully automatic mode to register the main most informative human bio-parameters. Among them, in the first place are the parameters characterizing the current state of the cardiovascular system of man, his breathing system, as well as his peripheral nervous system. The bulk of information is obtained by processing in real time the thermal infrared image of a person’s face. Unfortunately, existing algorithms for distinguishing a person’s face have a sufficiently high computational complexity and insufficient reliability. A typical example in this regard can be a family of algorithms based on the viola-Jones approach. The approach proposed in the work is based on taking into account additional information about the most likely location of a person’s face on a thermal image. This approach is most appropriate to use in cases of quasi-stationary location of people in the room. A typical example is the location of students at the tables in the classroom. For such cases it is possible to determine the areas of the most probable location of the trainees’ faces, as well as the possible boundaries of their movement. Laboratory tests of the developed program on the basis of the proposed algorithm have confirmed its high productivity, as well as efficiency in identifying students faces in the classroom.
In this work we propose a fully end-to-end approach for multi-spectral image registration and fusion. Our fusion method combines images from different spectral channels into a single fused image using approaches for l...
详细信息
In this work we propose a fully end-to-end approach for multi-spectral image registration and fusion. Our fusion method combines images from different spectral channels into a single fused image using approaches for low and high frequency signals. A prerequisite of fusion is the geometric alignment between the spectral bands, commonly referred to as registration. Unfortunately, common methods for image registration of a single spectral channel might prove inaccurate on images from different modalities. For that end, we introduce a new algorithm for multi-spectral image registration, based on a novel edge descriptor of feature points. Our method achieves an accurate alignment allowing us to further fuse the images. It is experimentally shown to produce a high quality of multi-spectral image registration and fusion under challenging scenarios.
A key challenge when displaying and processing sensed real-time 3D data is efficiency of generating and post-processingalgorithms in order to acquire high quality 3D content. In contrast, our approach focuses on volu...
详细信息
A key challenge when displaying and processing sensed real-time 3D data is efficiency of generating and post-processingalgorithms in order to acquire high quality 3D content. In contrast, our approach focuses on volumetric generation and processingvolumetric data using an efficient low-cost hardware setting. Acquisition of volumetric data is performed by connecting several Kinect v2 scanners to a single PC that are subsequently calibrated using planar pattern. This process is by no means trivial and requires well designed algorithms for fast processing and quick rendering of volumetric data. This can be achieved by fusing efficient filtering methods such as Weighted median filter (WM), Radius outlier removal (ROR) and Laplace-based smoothing algorithm. In this context, we demonstrate the robustness and efficiency of our technique by sensing several scenes.
The transformation of the development of the Arctic is due to modern robotic systems. The use of unmanned vehicles in many industries in the Arctic provides an array of photo and video information. Accuracy of image a...
The transformation of the development of the Arctic is due to modern robotic systems. The use of unmanned vehicles in many industries in the Arctic provides an array of photo and video information. Accuracy of image analysis and pattern recognition is enhanced by image preprocessing. However, the existing binarization algorithms are not universal for images with different distortions and loss of information. The accuracy of binarization algorithms depends on many factors, such as shadows, uneven lighting, low contrast, noise, etc. images with different characteristics of light and noise are simulated in order to model various lighting conditions on information from digital cameras of robotic systems. The paper investigates global and adaptive image binarization algorithms. The binarized images were obtained using these algorithms and the results of binarized images recognition are compared by an optical character recognition system. An analysis of the comparison results showed that for images made in poor lighting conditions or images with low contrast, or images with high noise levels, adaptive binarization algorithms are better suited. However, in most cases it is not possible to obtain fully correctly recognized *** paper proposes a new binarization method based on artificial neural networks. The process of creating an artificial neural network is shown, include the parameters for determining the class of a pixel, the adjustment of weights, the architecture of an artificial neural network. A comparison of the proposed artificial neural network with existing image binarization algorithms demonstrate that in most cases the artificial neural network has the result of imageprocessing at the level of adaptive algorithms or higher. The proposed method of images binarization based on the image color characteristics analysis allows to solve image recognition tasks by robotized systems.
暂无评论