This paper presents an Application Specific Instruction Set Processor (ASIP) pruned for high-throughput and variable-length Fast Fourier Transform (FFT), which is a key component of various Orthogonal Frequency Divisi...
详细信息
This paper presents an Application Specific Instruction Set Processor (ASIP) pruned for high-throughput and variable-length Fast Fourier Transform (FFT), which is a key component of various Orthogonal Frequency Division Multiplexing (OFDM)-based wireless communication standards. The ASIP executes dedicated FFT instructions to process two radix-4 or four radix-2 butterfly operations every clock cycle. Furthermore, a shuffle-embedded register file and a programmable memory access coprocessor are employed to tackle the memory access bottleneck and reduce power consumption. The implementation results show that our ASIP requires only 892 clock cycles for a 1024-point FFT, which outperforms TI TMS320C64x DSP and Tensilica ConnX ASIP by 6.74X and 2.03X, respectively. A test chip of the proposed ASIP was fabricated using CMOS 65nm process with the core area of 1.9mm 2 . It consumes 85mW when it runs at the maximum frequency of 150MHz.
In Java programs, it needs to use the information of the method type to resolve the virtual method dynamically, which restricts the performance greatly. Currently, the solution is mainly the technique of inline cachin...
详细信息
Railway freight safety inspection is an important component of railway transportation safety production. Traditional manual outdoor inspection methods are associated with issues such as high labor intensity, low effic...
Railway freight safety inspection is an important component of railway transportation safety production. Traditional manual outdoor inspection methods are associated with issues such as high labor intensity, low efficiency, and the potential for oversight. To address these problems, this paper proposes a real-time automated railway freight vehicle inspection method based on channel pruning, aiming to improve the detection efficiency of freight trains and alleviate the pressure of manual inspection. First, a YOLOv5s model is constructed, consisting of functional modules such as Focus, BottleneckCSP, and SPP. Subsequently, the model is compressed using a channel pruning technique, reducing its size and enabling deployment on resource-constrained devices like small-scale machines. Finally, the model is adjusted to achieve quick and precise rail freight truck detection. The experimental findings indicate that the pruned model reduces the number of model parameters by 91.9%, decreases the model size by 11.23 MB, and achieves a mAP of 98.14%, which is only 0.183% less than the model without pruning. To demonstrate that the suggested approach is preferable, a comparison is conducted with the YOLOv5x, YOLOv5n, and YOLOv5m algorithms. The comparison results demonstrate that the proposed method has significantly faster forward inference times than YOLOv5x, YOLOv5m, and YOLOv5n, with reductions of 95.6 ms, 22 ms, and 3.4 ms, respectively. The model size is also smaller than YOLOv5x, YOLOv5m, and YOLOv5n by 168.43 MB, 39.15 MB, and 2.34 MB, respectively. Moreover, the mAP is only 0.15% and 0.07% lower than YOLOv5x and YOLOv5m, respectively. These findings show that the suggested method, which can be implemented on small devices, achieves automated inspection of railway freight cars while taking detection speed and model size into account.
The advancement of large language models (LLMs) has significantly boosted the performance of complex long-form question answering tasks. However, one prominent issue of LLMs is the generated "hallucination" ...
详细信息
This paper presents a new method to detect naked people in an image using multiple features. The skin color model is firstly used to detect naked skin areas roughly. The Sobel edge operator and Gabor filter are used t...
详细信息
This paper presents a new method to detect naked people in an image using multiple features. The skin color model is firstly used to detect naked skin areas roughly. The Sobel edge operator and Gabor filter are used to weed those that are not really human skin pixels. Images that have many naked skin areas are thought maybe to have naked people. The color coherence vector and color histogram of these images are calculated and the SVM (support vector machine) is used to determine which of these images contain images of naked people.
In the industrial production process, the rolling bearing failures of huge mechanical equipment such as CNC machine tools frequently occur, which seriously affects the production performance and service life of the ma...
详细信息
This article aims to enhance the monitoring accuracy of high concurrent network services. As modern network services grow rapidly in data centers, tail latency has become one of the most crucial deciding factors on us...
详细信息
Hardware Trojan (HT) has paid more and more attention to the academia and industry because of its significant potential threat. In this paper, we propose a novel approach, named GramsDet, to detect HT through capturin...
详细信息
ISBN:
(数字)9781728126951
ISBN:
(纸本)9781728126968
Hardware Trojan (HT) has paid more and more attention to the academia and industry because of its significant potential threat. In this paper, we propose a novel approach, named GramsDet, to detect HT through capturing suspicious circuit connection structure using recurrent neural network. GramsDet considers that HT usually be inserted into the regions with low transition probability, so the circuit fragments associated with HT should have special connection structures. GramsDet models the target circuit using n-gram circuit segmentation technique, and implements the "gate embedding" by the order-sensitive co-occurrence matrix. Then, a stacked long short-term memory network is designed to build a robust HT detection model. The experimental results on different benchmarks show that GramsDet can detect effectively Trojan logic without the "Golden model" of the circuit under detection (CUD).
Recently, large language models (LLMs) have demonstrated excellent performance in understanding human instructions and generating code, which has inspired researchers to explore the feasibility of generating RTL code ...
详细信息
Existing approaches to improve the performances of convolutional neural networks by optimizing the local architectures or deepening the networks tend to increase the size of models significantly. In order to deploy an...
详细信息
ISBN:
(数字)9781728163956
ISBN:
(纸本)9781728163963
Existing approaches to improve the performances of convolutional neural networks by optimizing the local architectures or deepening the networks tend to increase the size of models significantly. In order to deploy and apply the neural networks to edge devices which are in great demand, reducing the scale of networks is quite crucial. However, It is easy to degrade the performance of image processing by compressing the networks. In this paper, we propose a method which is suitable for edge devices while improving the efficiency and effectiveness of inference. The joint decision of multiparticipants, mainly contain multi-layers and multi-networks, can achieve higher classification accuracy (0.26% on CFAR-10 and 4.49% on CFAR-100 at most) with similar total number of parameters for classical convolutional neural networks.
暂无评论