{"title":"I2CU: A Dedicated Im2col Hardware Unit","authors":"Tao Zhongyu, Wang Yuanfeng, Zhang Huaisheng","doi":"10.1109/iccwamtip56608.2022.10016515","DOIUrl":"https://doi.org/10.1109/iccwamtip56608.2022.10016515","url":null,"abstract":"For Convolution Neural Network (CNN), the convolution operation for feature map and weight map usually implemented by im2col + GEMM method. However, for conventional method need expand feature map to a large feature matrix during a single kernel function based on convolution parameters (i.e. filter size, padding, and stride), then multiplication for matrixes took place in another function. Thus the conventional method will generate tons data transfer and the large feature matrix requires enormous storage space, it is hardware unfriendly.We design a hardware unit, I2CU (Im2Col Unit), a dedicated hardware unit to implement im2col in hardware friendly way. I2CU dynamically expand loaded 4D-Block return from texture unit and write back destination matrix to shared memory. I2CU can decrease the feature matrix storage space and implement im2col + GEMM in one kernel function.","PeriodicalId":159508,"journal":{"name":"2022 19th International Computer Conference on Wavelet Active Media Technology and Information Processing (ICCWAMTIP)","volume":"73 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2022-12-16","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"126507615","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Chen Yuan, Yang Anjia, Liu Zilin, Pei Qingfu, Ma Li
{"title":"A Crosschain-Based Fair Exchange of Products For Poverty Alleviation","authors":"Chen Yuan, Yang Anjia, Liu Zilin, Pei Qingfu, Ma Li","doi":"10.1109/ICCWAMTIP56608.2022.10016552","DOIUrl":"https://doi.org/10.1109/ICCWAMTIP56608.2022.10016552","url":null,"abstract":"Through providing products or financial assistance, poverty alleviation is an effective way to reduce poverty. With properties of decentralization, traceability, and publicity, blockchain technology has been well studied to solve the problems of corruption, single point of failure, and complex transaction procedure in traditional centralized poverty alleviation platforms. However, most of poverty alleviation programs in the market are built on single blockchain, which contain only one kind of product information. It is difficult for products to be exchanged on different blockchains, which limits the further expansion of poverty alleviation. This paper proposes a universal and efficient cross-chain fair exchange scheme for poverty-alleviation products. Specifically, we employ an adaptor signature scheme to achieve fairness in the exchange. It ensures that when one party gets the exchanged poverty alleviation product, the other party will also get his corresponding product. In addition, we also adopt the verifiable time discrete logarithm algorithm to solve the problem that the lack of a special scripting language does not allow us to set \"timeout\" on the transaction and fall into a deadlock problem. Finally, we deploy the scheme on the local environment, public test chains of Ethereum, and conduct a series of experiments to verify its effectiveness and efficiency.","PeriodicalId":159508,"journal":{"name":"2022 19th International Computer Conference on Wavelet Active Media Technology and Information Processing (ICCWAMTIP)","volume":"6 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2022-12-16","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"127909105","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
{"title":"M3E-Yolo: A New Lightweight Network for Traffic Sign Recognition","authors":"Guo Haoran, Li Fan, Kuang Ping, Xiong Gang","doi":"10.1109/ICCWAMTIP56608.2022.10016618","DOIUrl":"https://doi.org/10.1109/ICCWAMTIP56608.2022.10016618","url":null,"abstract":"Traffic sign recognition is committed to ensuring the safety of automatic driving. Inspired by YOLOv5, this paper proposes a new model to solve the problem of poor balance between the accuracy and efficiency of existing algorithms in traffic sign recognition. Firstly, the lightweight network MobileNetV3 is introduced for feature extraction to reduce the number of parameters. Secondly, attention mechanism module is introduced to enhance channel features, which makes up for the reduced accuracy caused by the simplified model. Experiments show that the mAP value trained by our model on the Chinese traffic sign dataset reaches 93.6%, which is similar to the level of YOLOv5, and the number of parameters is less than a quarter of YOLOv5.","PeriodicalId":159508,"journal":{"name":"2022 19th International Computer Conference on Wavelet Active Media Technology and Information Processing (ICCWAMTIP)","volume":"49 s171","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2022-12-16","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"132226856","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
Lin Yuzhou, Wu Xiaofei, Lin Mao, Lu Chao, Wan Hu, Chen Yongqiang, Sun Jingbo
{"title":"A Hybrid Model of Kalman-ARIMA-LSTM for Flow Prediction of Mine air Compressors","authors":"Lin Yuzhou, Wu Xiaofei, Lin Mao, Lu Chao, Wan Hu, Chen Yongqiang, Sun Jingbo","doi":"10.1109/ICCWAMTIP56608.2022.10016570","DOIUrl":"https://doi.org/10.1109/ICCWAMTIP56608.2022.10016570","url":null,"abstract":"As an essential power system in coal mine production, the accidental failure or downtime of mine air compressors will cause production stagnation and economic losses. The accurate and reliable flow prediction of air compressors plays a crucial role in fault diagnosis for air compressors. In this paper, we propose a hybrid model combining the Kalman filter, the autoregressive integrated moving average and the long short-term memory to predict the flow data of mine air compressors. The noise existing in the flow data of air compressors can be reduced by the Kalman filter so that the sequence can be recognized by the autoregressive integrated moving average model. Then, the long short-term memory neural network will be utilized to transform the linear data to nonlinear for conforming to the actual flow situation of air compressors. The prediction results show that the hybrid model has a good performance in forecasting the flow data of mine air compressors.","PeriodicalId":159508,"journal":{"name":"2022 19th International Computer Conference on Wavelet Active Media Technology and Information Processing (ICCWAMTIP)","volume":"1 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2022-12-16","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"130285250","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
{"title":"Optimization of Sequencing Problem for Automobile Manufacturing Based on Sorting Buffer","authors":"Zhang Ruizhi, L. Jun, S. Xiaofeng","doi":"10.1109/ICCWAMTIP56608.2022.10016520","DOIUrl":"https://doi.org/10.1109/ICCWAMTIP56608.2022.10016520","url":null,"abstract":"Painted Body Store(PBS) buffer can change the loading sequence of the vehicles, which can meet the requirements of the subsequent production requirements. In this paper, an optimization model of in-and-out vehicle scheduling in PBS is considered, and the vehicle scheduling at the entrance and exit of the buffer is realized respectively. At the entrance of the delivery lane, a heuristic rule is set up to select the optimal lane according to the congestion degree of each available. At the exit of the delivery lane, based on the greedy criterion, it determines whether the vehicle should be taken out, or sent back to the return lane for re-scheduling. In addition, to avoid getting caught in loop scheduling, the penalty term is given to the number of allowed uses of the return path according to the scheduling time cost at the current moment, so that the performance and time cost can be balanced to some extent.","PeriodicalId":159508,"journal":{"name":"2022 19th International Computer Conference on Wavelet Active Media Technology and Information Processing (ICCWAMTIP)","volume":"15 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2022-12-16","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"132844974","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
{"title":"An Adaptive Hardware Accelerator For Convolution Layers With Diverse Sizes","authors":"Zhao Yiwei, Zou Hao, Tang Ming, Lin Qiutong","doi":"10.1109/ICCWAMTIP56608.2022.10016562","DOIUrl":"https://doi.org/10.1109/ICCWAMTIP56608.2022.10016562","url":null,"abstract":"Convolution is the most important operation in convolutional neural networks (CNN). FPGA-based CNN accelerators need to fully consider the optimization of convolution loops to get ideal performance. This work analyzes convolution loop optimization in detail, exploiting loop tiling, loop unrolling, and loop interchange to design the dataflow of accelerator. This work quantitatively evaluates strategies for data reuse and resource utilization, combining fixed and dynamic parallelism to design a high-performance adaptive accelerator. The proposed accelerator is evaluated on ZCU102 FPGA by implementing a five-layer CNN with large differences in convolution layer sizes. It achieves more than 1.14x improvement in throughput efficiency over prior accelerators. And the consumption of logic resources is less than half of prior accelerators while the computing resources are similar.","PeriodicalId":159508,"journal":{"name":"2022 19th International Computer Conference on Wavelet Active Media Technology and Information Processing (ICCWAMTIP)","volume":"9 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2022-12-16","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"131869375","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
You Jiang, Gu Jingliang, Z. Yanqing, W. Min, Wang Jianwei
{"title":"Detection and Tracking Method of Small-Sized UAV Based on YOLOv5","authors":"You Jiang, Gu Jingliang, Z. Yanqing, W. Min, Wang Jianwei","doi":"10.1109/ICCWAMTIP56608.2022.10016550","DOIUrl":"https://doi.org/10.1109/ICCWAMTIP56608.2022.10016550","url":null,"abstract":"In recent years, accidents such as casualties and economic losses have occurred due to unmanned aerial vehicles (UAVs) flying in violation of regulations. The monitoring and countermeasures of UAVs with low flying altitude, slow flying speed and small size have become the current research hotspot. Vision-based methods are still one of the most mainstream methods, but due to the characteristics of UAVs such as small size, large attitude changes, and low flight altitudes, there are difficulties such as poor imaging contrast, complex background, and small proportion of targets. Aiming at the above difficulties, this paper proposes an improved YOLOv5 UAV detection algorithm and tracking method. The detection probability of drones is improved by adding a detection heads and attention module, and high-speed tracking performance is achieved by training a low-resolution detector combined with the Kalman algorithm. This paper deploys this method on NVIDIA Jetson Xavier NX, resulting in an output rate of 200 FPS.","PeriodicalId":159508,"journal":{"name":"2022 19th International Computer Conference on Wavelet Active Media Technology and Information Processing (ICCWAMTIP)","volume":"28 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2022-12-16","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"133726712","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
{"title":"Design and Implementation of Beidou-Based Intelligent Vehicle Terminal","authors":"Li Shiyou","doi":"10.1109/ICCWAMTIP56608.2022.10016564","DOIUrl":"https://doi.org/10.1109/ICCWAMTIP56608.2022.10016564","url":null,"abstract":"Intelligent vehicle terminal can collect the message of position, speed, temperature, and other information by kinds of sensors and then monitor the real-time status of vehicle for driver or managing center. It is an important component of intelligent transportation system and has an important impact on the management of traffic. In the current report, a Beidou-based intelligent vehicle terminal is designed and implemented, which can collect vehicle information in real time and communicate with managing center through General Packet Radio Service (GPRS) wireless network. The hardware circuit of the vehicle terminal is compose of a main control module with STM32 as the processor, a Beidou navigation and positioning module, a GPRS wireless communication module, and a temperature sensor module. The system is successfully implemented after finishing software developing and system testing. Experimental results show that the vehicle terminal system based on Beidou satellite navigation and positioning technology can operate normally and stably, and can locate the vehicle in real time on the mobile phone application software. The system can be applied such as fleet operation status management.","PeriodicalId":159508,"journal":{"name":"2022 19th International Computer Conference on Wavelet Active Media Technology and Information Processing (ICCWAMTIP)","volume":"27 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2022-12-16","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"115571496","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
{"title":"Road Flat Condition Detection Algorithm Based on Group Intelligence Perception","authors":"Chen Yang, Xing Jianchuan, Zhou Chunwen","doi":"10.1109/ICCWAMTIP56608.2022.10016553","DOIUrl":"https://doi.org/10.1109/ICCWAMTIP56608.2022.10016553","url":null,"abstract":"Road traffic is the lifeblood of national economic development. Although all kinds of transportation modes are developing very rapidly, road traffic is still irreplaceable. Because of China’s vast territory, there are too many areas that trains, planes, ships and other vehicles cannot reach, and then the road traffic becomes the only means of transportation. So it is very important to maintain the road health in real time. However, the current way to obtain road health conditions mainly depends on the relevant departments to spend a lot of material resources and the use of professional measuring instruments to measure, there is a problem of not timely access to road conditions. In this paper, we combine the same group intelligence perception technology, use the sensor carried by the intelligent terminal to collect the vehicle driving data, and design the algorithm to process the collected data and then get the result of the road flatness.","PeriodicalId":159508,"journal":{"name":"2022 19th International Computer Conference on Wavelet Active Media Technology and Information Processing (ICCWAMTIP)","volume":"1 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2022-12-16","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"130071393","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
{"title":"RETR: End-To-End Referring Expression Comprehension with Transformers","authors":"Yang Rui","doi":"10.1109/ICCWAMTIP56608.2022.10016599","DOIUrl":"https://doi.org/10.1109/ICCWAMTIP56608.2022.10016599","url":null,"abstract":"Referring Expression Comprehension (REC) is a basic and challenging task to identify the referred region given a language expression. However, existing two-stage or one-stage methods suffer from the region proposals, the limited range of visual context and the incomplete cross-modal alignment. To address these problems, we propose a simple yet effective one-stage model, termed REC TRansformer (RETR), which is trained end-to-end. Different from the manually designed multi-modal fusion, RETR adopts a transformer decoder with alternately stacked self-attention and cross-attention layers to capture the global visual context and establish the detailed visual-linguistic correspondence. Moreover, we utilize multiple learnable tokens to obtain diverse yet complementary region representations to give the accurate prediction. Extensive experiments are conducted on four datasets and RETR achieves the state-of-the-art performance.","PeriodicalId":159508,"journal":{"name":"2022 19th International Computer Conference on Wavelet Active Media Technology and Information Processing (ICCWAMTIP)","volume":"19 1","pages":"0"},"PeriodicalIF":0.0,"publicationDate":"2022-12-16","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"126491181","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":0,"RegionCategory":"","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}