Sophisticated computational imaging algorithms require both high performance and good energy-efficiency when executed on mobile devices. Recent trend has been to exploit the abundant data-level parallelism found in ge...
详细信息
ISBN:
(纸本)9781509010950
Sophisticated computational imaging algorithms require both high performance and good energy-efficiency when executed on mobile devices. Recent trend has been to exploit the abundant data-level parallelism found in general purpose programmable GPUs. However, for low-power mobile use cases, generic GPUs consume excessive amounts of power. This paper proposes a programmable computational imaging processor with 16-bit half-precision SIMD floating point vector processing capabilities combined with power efficiency of an exposed datapath. In comparison to traditional VLIW architectures with similar computational resources, the exposed datapath reduces the register file traffic and complexity. These and the specific optimizations enabled by the explicit programming model enable extremely good power-performance. When synthesized on a 28nm ASIC technology, the accelerator consumes 71mW of power while running a state-of-the-art denoising algorithm, and occupies only 0.2mm(2) of chip area. For the algorithm, energy usage per frame is 7mJ, which is 10x less than the best found GPU-based implementation.
This paper proposes an efficient method to improve image quality based on Context-based enhancement techniques, particularly towards real-time applications in dedicated hardware systems. The main idea is that all the ...
详细信息
ISBN:
(纸本)9781467380171
This paper proposes an efficient method to improve image quality based on Context-based enhancement techniques, particularly towards real-time applications in dedicated hardware systems. The main idea is that all the important information of a low-quality an image are combined with a high-quality daytime background image at the same scene to output an image with improved quality, and thus main objects of the low-quality input images can be recognized more easily. In comparison, almost all current context-based enhancement techniques have processed with complex algorithms, thereby being time-consuming and hard to perform in real-time processing when applying into high-resolution images/videos. The developed technique here outperforms these original methods, which is demonstrated by without complex formulas (such as DST and IDST), whereas not only still keeping all the important information of the low-quality nighttime input images, but also limiting problems often existing in context fusion techniques, such as ghosting, haloing and artifact effects. The experimental results demonstrate that our approach has been efficient and gained a better quality of output images in comparison with the results of the Denighting method [8].
Automatic Identification of disease through imageprocessing in biomedical field is the norm of modern era. Ophthalmologists have used several invasive and noninvasive techniques for early detection of disease. OCT is...
详细信息
ISBN:
(纸本)9781509018185
Automatic Identification of disease through imageprocessing in biomedical field is the norm of modern era. Ophthalmologists have used several invasive and noninvasive techniques for early detection of disease. OCT is one such noninvasive modality that performs high resolution tomographic imaging in biological systems. OCT images are produced containing speckle noise. Noise is a major factor that decreases image quality, hence degrading performance of noise imageprocessingalgorithms. It is therefore at highest priority to apply effective method for denoising image, before further processing. In this paper we applied Wavelet denoising, bilateral and wiener filter on an OCT images and discuss benefits and drawback of each algorithms. The effectiveness of each algorithm is compared on basis of Signal to noise ratio SNR, peak signal to noise ratio PSNR and Mean square error MSE.
Automatic face recognition technologies have seen significant improvements in performance due to a combination of advances in deep learning and availability of larger datasets for training deep networks. Since recogni...
详细信息
ISBN:
(纸本)9781509014378
Automatic face recognition technologies have seen significant improvements in performance due to a combination of advances in deep learning and availability of larger datasets for training deep networks. Since recognizing faces is a task that humans are believed to be very good at, it is only natural to compare the relative performance of automated face recognition and humans when processing fully unconstrained facial imagery. In this work, we expand on previous studies of the recognition accuracy of humans and automated systems by performing several novel analyses utilizing unconstrained face imagery. We examine the impact on performance when human recognizers are presented with varying amounts of imagery per subject, immutable attributes such as gender, and circumstantial attributes such as occlusion, illumination, and pose. Results indicate that humans greatly outperform state of the art automated face recognition algorithms on the challenging IJB-A dataset.
Multi-beam scanning electron microscopy (MBSEM), has been developed to reduce the acquisition time by scanning multiple pixels simultaneously. The signal from the 14 × 14 beams is captured on a camera which reads...
详细信息
ISBN:
(纸本)9781467395564
Multi-beam scanning electron microscopy (MBSEM), has been developed to reduce the acquisition time by scanning multiple pixels simultaneously. The signal from the 14 × 14 beams is captured on a camera which reads out the position and intensity for each beam on the sample. But as we work with multiple beams and pixels we need a powerful technique for image acquisition and imageprocessingalgorithms. We use Field Programmable Gate Arrays (FPGA's), often used as an implementation platform for real time image acquisition and processing applications, because their structure is able to exploit spatial and temporal parallelism. This paper presents a technique for dealing with the various constraints of the camera and efficient mapping for image acquisition and processing operations on FPGA.
Temporal psychovisual modulation (TPVM) is a new information display technology which aims to generate multiple visual percepts for different viewers on a single display simultaneously. In a TPVM system, the viewers w...
详细信息
Todays, the number of vehicles is rapidly increasing. In parallel, the number of ways and traffic signs have increased. As a result of increased traffic signs, the drivers are expected to learn all the traffic signs a...
详细信息
Todays, the number of vehicles is rapidly increasing. In parallel, the number of ways and traffic signs have increased. As a result of increased traffic signs, the drivers are expected to learn all the traffic signs and to pay attention to them while driving. A system that can automatically recognize the traffic signs has been need to reduce traffic accidents and to drive more freely. Traffic sign recognition system meet this need. This study includes traffic sign detection and recognition application. In this study, some imageprocessing techniques are used to detect traffic signs and Fuzzy Integral is used to recognize traffic signs. Both more accuracy rate results and low computational cost are obtained in terms of recognition stage by using positive aspects of algorithms taken as input parameters with Fuzzy Integral in the traffic sign recognition system. Experimental results show that proposed method gives high accurate results in a reasonable time.
In this paper, imageprocessingalgorithms designed in Zynq SoC using the Vivado HLS tool are presented and compared with hand-coded designs. In Vivado HLS, the designer has the opportunity to employ libraries similar...
详细信息
ISBN:
(纸本)9781509045662
In this paper, imageprocessingalgorithms designed in Zynq SoC using the Vivado HLS tool are presented and compared with hand-coded designs. In Vivado HLS, the designer has the opportunity to employ libraries similar to OpenCV, a library that is well-known and wide used by software designers. The algorithms are compared in terms of area resources in two conditions: using the libraries and not using the libraries. The case studies are Data Binning, a Step Row Filter and a Sobel Filter. These algorithms have been selected because they are very common in the field of imageprocessing and they have high computational complexity. The main benefit of the Vivado HLS tool is the reduction in time-to-market. On the other hand, when a software designer hand-codes the design, the use of imageprocessing libraries similar to OpenCV helps to reduce development time even further because software designers are familiar with them. However, using these kinds of libraries significantly increases the necessary FPGA resources.
Summary form only given. Over the recent years the role of mathematics in innovations for Circuits, systems and Signal processing has increased considerably. The talk will overview the dynamical forces and their impac...
详细信息
Summary form only given. Over the recent years the role of mathematics in innovations for Circuits, systems and Signal processing has increased considerably. The talk will overview the dynamical forces and their impact on the research and education. Examples will be given of mathematical methodologies for signal and image classification, data fusion, biomedical diagnostics with support vector machines, matrix and tensor decompositions. Also cryptographic algorithms are crucial in our modern society. Important lessons can be learned for research planning, dissemination and reproducibility as well as for teaching in engineering.
Imaging systems have been applied to many new applications in recent years. With the advent of low-cost, low-power focal planes and more powerful, lower cost computers, remote sensing applications have become more wid...
Imaging systems have been applied to many new applications in recent years. With the advent of low-cost, low-power focal planes and more powerful, lower cost computers, remote sensing applications have become more wide spread. Many of these applications require some form of geolocation, especially when relative distances are desired. However, when greater global positional accuracy is needed, orthorectification becomes necessary. Orthorectification is the process of projecting an image onto a Digital Elevation Map (DEM), which removes terrain distortions and corrects the perspective distortion by changing the viewing angle to be perpendicular to the projection plane. Orthorectification is used in disaster tracking, landscape management, wildlife monitoring and many other applications. However, orthorectification is a computationally expensive process due to floating point operations and divisions in the algorithm. To reduce the computational cost of on-board processing, two novel algorithm modifications are proposed. One modification is projection utilizing fixed-point arithmetic. Fixed point arithmetic removes the floating point operations and reduces the processing time by operating only on integers. The second modification is replacement of the division inherent in projection with a multiplication of the inverse. The inverse must operate iteratively. Therefore, the inverse is replaced with a linear approximation. As a result of these modifications, the processing time of projection is reduced by a factor of 1.3x with an average pixel position error of 0.2% of a pixel size for 128-bit integer processing and over 4x with an average pixel position error of less than 13% of a pixel size for a 64-bit integer processing. A secondary inverse function approximation is also developed that replaces the linear approximation with a quadratic. The quadratic approximation produces a more accurate approximation of the inverse, allowing for an integer multiplication calculation
暂无评论