Different from binary computation, stochastic computation (SC), as a new paradigm, uses stochastic bit stream (SBS) to encode data. By simplifying computing elements, the circuit area can be greatly reduced. SBS can b...
详细信息
ISBN:
(纸本)9781665450737
Different from binary computation, stochastic computation (SC), as a new paradigm, uses stochastic bit stream (SBS) to encode data. By simplifying computing elements, the circuit area can be greatly reduced. SBS can be generated by a stochastic number generator (SNG) with a variety of formats. In this work, we use unipolar (UP) and bipolar (BP) formats to optimize the traditional SC subtractor, which is named the UP-to-BP Subtractor (UBS). A new cross format coding (CFC) method is proposed for stochastic computing, which combines the UP and BP format, and is applied to Sobel edge detection in imageprocessingalgorithms. The fault tolerance and detection efficacy of the proposed CFC method and conventional binary computing are compared in this paper. By using the CFC method, the detected F-Score is improved by 0.15(23%). If the F-score remains unchanged, the processing speed can be about 10 times faster.
image detection plays a vital role in digital imageprocessing and artificial intelligence, with applications ranging from security surveillance to autonomous vehicles and medical image analysis. This study employs a ...
详细信息
ISBN:
(数字)9798331517601
ISBN:
(纸本)9798331517618
image detection plays a vital role in digital imageprocessing and artificial intelligence, with applications ranging from security surveillance to autonomous vehicles and medical image analysis. This study employs a qualitative approach to review the performance of image detection algorithms, focusing on deep learning methods such as YOLO, SSD, and CNN. Through computational experiments using quantitative data, this research provides performance comparisons based on accuracy, precision, recall, and computational efficiency. The results highlight YOLO’s superior performance in terms of both accuracy ($\mathbf{9 2} .5 \%$) and inference speed (75.2 fps), making it suitable for real-time applications. This study contributes by addressing the gap in balancing computational efficiency and adaptability for real-world applications.
China's new rural construction project has developed rapidly. The rural landscape has changed a lot. The scale of the rural roads is getting larger and larger. The maintenance cost of rural road diseases is also v...
详细信息
Data compression is a trending field that is used in data storage and data transmission systems. Lossy compression means that data cannot be completely retrieved while in lossless compression the compressed data must ...
详细信息
ISBN:
(纸本)9781665482370
Data compression is a trending field that is used in data storage and data transmission systems. Lossy compression means that data cannot be completely retrieved while in lossless compression the compressed data must be reconstructed exactly. Lossless data compression is used in compressing binary files, telemetry data and high-fidelity medical and scientific images where details are crucial. There is no generic compression algorithm that gives best compression ratio on all data pattern. In this paper, we propose a hybrid lossless hardware architecture that compresses most of data patterns such as repeated data, Gaussian distribution data and images. A profiling-before-compressing and then choosing the right compression hardware is proposed. The proposed design is a highly parallelized architecture that can compress/decompress 64 bytes/cycle with minor overhead. Moreover, it provides high compression ratio on small block sizes as well as large ones.
images captured in surveillance systems suffer from low contrast and faint color. Recently, plenty of dehazing algorithms have been proposed to enhance visibility and restore color. We present a new image enhancement ...
详细信息
ISBN:
(纸本)9781510655546
images captured in surveillance systems suffer from low contrast and faint color. Recently, plenty of dehazing algorithms have been proposed to enhance visibility and restore color. We present a new image enhancement algorithm based on multi-scale block-rooting processing. The basic idea is to apply the frequency domain image enhancement approach for different image block scales. The parameter of transform coefficient enhancement for every block is driven through optimization of measure of enhancement. The main idea is that enhancing the contrast of an image would create more high-frequency content in the enhanced image than the original image. To test the performance of the proposed algorithm, the public database O-HAZE is used.
This study focuses on dust detection in solar panels by utilizing imageprocessing techniques. Dust detection helps in forecasting the maintenance needs and ensures system reliability. Inefficient maintenance practice...
详细信息
ISBN:
(数字)9798350378177
ISBN:
(纸本)9798350378184
This study focuses on dust detection in solar panels by utilizing imageprocessing techniques. Dust detection helps in forecasting the maintenance needs and ensures system reliability. Inefficient maintenance practices of solar panels lead to increased downtime and decreased energy production. Implementation of dust detection holds importance in ensuring consistent power output. In the existing methodologies, Artificial Neural Network algorithms are implemented for dust detection. Artificial Neural Network algorithms show a considerably lower accuracy in imageprocessing due to overfitting and complexity of image data. In the proposed system, a comparative analysis is performed by implementing Multi-Layer Perceptron technique and Dense Net algorithm. Using Multi-Layer Perceptron, a Sequential model is built with one flatten layer and two dense layers. It shows an accuracy of 88%. The Dense Net algorithm was also implemented on the processed image dataset and it shows an accuracy of 98%. This helps in increasing the accuracy and minimizing error. The Thonny IDE is employed for the implementation of the models, with validation conducted using hardware components such as Raspberry Pi, solar panels, and camera modules. Real-time images are captured to assess the performance. The utilization of the Convolution Neural Network algorithm, specifically Dense Net, for dust detection has led to enhanced system accuracy and efficiency.
The spotlight on large population-based studies is growing within the research community. Epidemiological studies, amassing extensive data through questionnaires and check-ups, often include imaging data like magnetic...
详细信息
Implementing image dehazing and defogging on a Field Programmable Gate Array (FPGA) offers efficiency. Dehazing an image becomes particularly challenging in the presence of fog or haze. However, employing a dark chann...
详细信息
ISBN:
(数字)9798350382693
ISBN:
(纸本)9798350382709
Implementing image dehazing and defogging on a Field Programmable Gate Array (FPGA) offers efficiency. Dehazing an image becomes particularly challenging in the presence of fog or haze. However, employing a dark channel prior to dehazing allows the removal of haze particles from the image. Different modules are used in this process, such as Dynamic Atmospheric Light Estimation (DALE), Scene Recovery (SR), and Transmission Map Estimation (TME). The FPGA runs these modules in hardware and produces effective outcomes by employing imageprocessingalgorithms. In this case, using FPGA technology offers a number of benefits. The dehazing process can be accelerated by using FPGA's built-in parallel processing capabilities to execute numerous operations at once. Furthermore, FPGA implementations provide better throughput and reduced latency in comparison to conventional approaches, making them well-suited for real-time applications such as image dehazing and D
Eye diseases have always been a threat to public health worldwide. Many people suffer from various eye diseases, but there are not enough skilled ophthalmologists to meet the demand for medical care. Thus, finding a m...
详细信息
ISBN:
(纸本)9781510666313;9781510666320
Eye diseases have always been a threat to public health worldwide. Many people suffer from various eye diseases, but there are not enough skilled ophthalmologists to meet the demand for medical care. Thus, finding a method to perform ophthalmic examinations automatically and conveniently is necessary. Although many well-designed ophthalmic diagnosis systems have been proposed to diagnose ophthalmic disorders using artificial intelligence algorithms, they tend to depend on high-quality anterior segment images to perform appropriately. In order to capture high-quality anterior segment images simply with a smartphone, we proposed a system including a semantic segmentation model and an image quality assessment method for anterior segment images. Our proposed segmentation model, namely the multi-task anterior segment image semantic segmentation (MT-ASISS) model, has a designed multi-task learning network structure and achieves an accuracy of 92.63% in Dice and a processing speed of 138ms per frame on smartphones. Our anterior segment image quality assessment method, namely Mixed-Parameters Quality Assessment (MPQA) method, has an accuracy of 92.6% in mean average precision (mAP). The system can help reduce the demand for professional image collecting equipment, share the burden of choosing satisfactory images manually and improve the efficiency of acquiring anterior segment images.
Due to the complexity of their structure and the particularity of their application environments, aircraft High-Voltage Direct Current (HVDC) systems are prone to faults, with inter-module failures complicating fault ...
详细信息
ISBN:
(数字)9798331541460
ISBN:
(纸本)9798331541477
Due to the complexity of their structure and the particularity of their application environments, aircraft High-Voltage Direct Current (HVDC) systems are prone to faults, with inter-module failures complicating fault diagnosis. To address this issue, a fault diagnosis method for HVDC systems has been developed. Feature extraction methods were designed for the rectifier, BUCK converter, and inverter, respectively, with the sum-to-amplitude ratio of signals selected as a feature for the rectifier; multi-scale skewness was proposed for the BUCK converter; and the ratio of the signal's average to peak absolute value was chosen for the inverter. Subsequently, the PSO-LightGBM algorithm was proposed, which employs the LightGBM algorithm for classification and utilizes a particle swarm algorithm to optimize the parameters of the LightGBM, establishing the optimal model. The experimental results demonstrate that the proposed method can accurately achieve fault diagnosis in HVDC systems.
暂无评论