检索结果-内蒙古大学图书馆

Signal propagation in transformers: theoretical perspectives and the role of rank collapse 22

学校读者我要写书评

暂无评论

Signal propagation in transformers: theoretical perspectives...

Proceedings of the 36th International Conference on Neural Information Processing Systems

作者： Lorenzo Noci Sotiris Anagnostidis Luca Biggio Antonio Orvieto Sidak Pal Singh Aurelien Lucchi Dept of Computer Science ETH Zürich Dept of Computer Science ETH Zürich and Robotics & ML CSEM SA Alpnach Switzerland Dept of Computer Science ETH Zürich and MPI for Intelligent Systems Tübingen Department of Mathematics and Computer Science University of Basel

ISBN: (纸本)9781713871088

Transformers have achieved remarkable success in several domains, ranging from natural language processing to computer vision. Nevertheless, it has been recently shown that stacking self-attention layers — the distinctive architectural component of Transformers — can result in rank collapse of the tokens' representations at initialization. The question of if and how rank collapse affects training is still largely unanswered, and its investigation is necessary for a more comprehensive understanding of this architecture. In this work, we shed new light on the causes and the effects of this phenomenon. First, we show that rank collapse of the tokens' representations hinders training by causing the gradients of the queries and keys to vanish at initialization. Furthermore, we provide a thorough description of the origin of rank collapse and discuss how to prevent it via an appropriate depth-dependent scaling of the residual branches. Finally, our analysis unveils that specific architectural hyperparameters affect the gradients of queries and values differently, leading to disproportionate gradient norms. This suggests an explanation for the widespread use of adaptive methods for Transformers' optimization.

关键词：

Fast-moving object counting with an event camera

学校读者我要写书评

暂无评论

TechRxiv

TechRxiv 2022年

作者： Bialik, Kamil Kowalczyk, Marcin Blachut, Krzysztof Kryjak, Tomasz Embedded Vision Systems Group Computer Vision Laboratory Department of Automatic Control and Robotics AGH University of Science and Technology Al. Mickiewicza 30 Krakow30-059 Poland

This paper proposes the use of an event camera as a component of a vision system that enables counting of fast-moving objects – in this case, falling corn grains. These type of cameras transmit information about the change in brightness of individual pixels and are characterised by low latency, no motion blur, correct operation in different lighting conditions, as well as very low power consumption. The proposed counting algorithm processes events in real time. The operation of the solution was demonstrated on a stand consisting of a chute with a vibrating feeder, which allowed the number of grains falling to be adjusted. The objective of the control system with a PID controller was to maintain a constant average number of falling objects. The proposed solution was subjected to a series of tests to determine the correctness of the developed method operation. On their basis, the validity of using an event camera to count small, fast-moving objects and the associated wide range of potential industrial applications can be confirmed. © 2022-12-16, CC BY.

关键词： Cameras

PointPillars Backbone Type Selection For Fast and Accurate LiDAR Object Detection

学校读者我要写书评

暂无评论

arXiv 2022年

关键词： Object detection

Skeletal Human Action Recognition using Hybrid Attention based Graph Convolutional Network

学校读者我要写书评

暂无评论

arXiv 2022年

作者： Xing, Hao Burschka, Darius Technical University of Munich Machine Vision and Perception Group Munich Institute of Robotics and Machine Intelligence Department of Computer Science Parkring 13 Munich85748 Germany

In skeleton-based action recognition, Graph Convolutional Networks model human skeletal joints as vertices and connect them through an adjacency matrix, which can be seen as a local attention mask. However, in most existing Graph Convolutional Networks, the local attention mask is defined based on natural connections of human skeleton joints and ignores the dynamic relations for example between head, hands and feet joints. In addition, the attention mechanism has been proven effective in Natural Language Processing and image description, which is rarely investigated in existing methods. In this work, we proposed a new adaptive spatial attention layer that extends local attention map to global based on relative distance and relative angle information. Moreover, we design a new initial graph adjacency matrix that connects head, hands and feet, which shows visible improvement in terms of action recognition accuracy. The proposed model is evaluated on two large-scale and challenging datasets in the field of human activities in daily life: NTU-RGB+D and Kinetics skeleton. The results demonstrate that our model has strong performance on both dataset. © 2022, CC BY.

关键词： Convolution

Traffic Sign Detection With Event Cameras and DCNN

学校读者我要写书评

暂无评论

arXiv 2022年

作者： Wzorek, Piotr Kryjak, Tomasz Embedded Vision Systems Group Computer Vision Laboratory Department of Automatic Control and Robotics AGH University of Science and Technology Kraków Poland Department of Digital Systems Silesian University of Technology Gliwice Poland

In recent years, event cameras (DVS – Dynamic vision Sensors) have been used in vision systems as an alternative or supplement to traditional cameras. They are characterised by high dynamic range, high temporal resolution, low latency, and reliable performance in limited lighting conditions – parameters that are particularly important in the context of advanced driver assistance systems (ADAS) and self-driving cars. In this work, we test whether these rather novel sensors can be applied to the popular task of traffic sign detection. To this end, we analyse different representations of the event data: event frame, event frequency, and the exponentially decaying time surface, and apply video frame reconstruction using a deep neural network called FireNet. We use the deep convolutional neural network YOLOv4 as a detector. For particular representations, we obtain a detection accuracy in the range of 86.9-88.9% mAP@0.5. The use of a fusion of the considered representations allows us to obtain a detector with higher accuracy of 89.9% mAP@0.5. In comparison, the detector for the frames reconstructed with FireNet is characterised by an accuracy of 72.67% mAP@0.5. The results obtained illustrate the potential of event cameras in automotive applications, either as standalone sensors or in close cooperation with typical frame-based cameras. © 2022, CC BY.

关键词： Cameras

Optimisation of a Siamese Neural Network for Real-Time Energy Efficient Object Tracking 1

学校读者我要写书评

暂无评论

International Conference on computer vision and Graphics, ICCVG 2020

作者： Przewlocka, Dominika Wasala, Mateusz Szolc, Hubert Blachut, Krzysztof Kryjak, Tomasz Embedded Vision Systems Group Computer Vision Laboratory Department of Automatic Control and Robotics AGH University of Science and Technology Krakow Poland

ISBN: (数字)9783030590062

ISBN: (纸本)9783030590055

In this paper the research on optimisation of visual object tracking using a Siamese neural network for embedded vision systems is presented. It was assumed that the solution shall operate in real-time, preferably for a high resolution video stream, with the lowest possible energy consumption. To meet these requirements, techniques such as the reduction of computational precision and pruning were considered. Brevitas, a tool dedicated for optimisation and quantisation of neural networks for FPGA implementation, was used. A number of training scenarios were tested with varying levels of optimisations – from integer uniform quantisation with 16 bits to ternary and binary networks. Next, the influence of these optimisations on the tracking performance was evaluated. It was possible to reduce the size of the convolutional filters up to 10 times in relation to the original network. The obtained results indicate that using quantisation can significantly reduce the memory and computational complexity of the proposed network while still enabling precise tracking, thus allow to use it in embedded vision systems. Moreover, quantisation of weights positively affects the network training by decreasing overfitting. © 2020, Springer Nature Switzerland AG.

关键词： Field programmable gate arrays (FPGA)

A vision Based Hardware-Software Real-Time Control System for the Autonomous Landing of an UAV 1

学校读者我要写书评

暂无评论

International Conference on computer vision and Graphics, ICCVG 2020

作者： Blachut, Krzysztof Szolc, Hubert Wasala, Mateusz Kryjak, Tomasz Gorgon, Marek Embedded Vision Systems Group Computer Vision Laboratory Department of Automatic Control and Robotics AGH University of Science and Technology Krakow Poland

ISBN: (数字)9783030590062

ISBN: (纸本)9783030590055

In this paper we present a vision based hardware-software control system enabling the autonomous landing of a multirotor unmanned aerial vehicle (UAV). It allows for the detection of a marked landing pad in real-time for a 1280 720 @ 60 fps video stream. In addition, a LiDAR sensor is used to measure the altitude above ground. A heterogeneous Zynq SoC device is used as the computing platform. The solution was tested on a number of sequences and the landing pad was detected with 96% accuracy. This research shows that a reprogrammable heterogeneous computing system is a good solution for UAVs because it enables real-time data stream processing with relatively low energy consumption. © 2020, Springer Nature Switzerland AG.

关键词： System-on-chip

Learning Multi-dimensional Edge Feature-based AU Relation Graph for Facial Action Unit Recognition

学校读者我要写书评

暂无评论

arXiv 2022年

作者： Luo, Cheng Song, Siyang Xie, Weicheng Shen, Linlin Gunes, Hatice Computer Vision Institute Shenzhen University China Shenzhen Institute of Artificial Intelligence and Robotics for Society China Guangdong Key Laboratory of Intelligent Information Processing China Department of Computer Science and Technology University of Cambridge United Kingdom

The activations of Facial Action Units (AUs) mutually influence one another. While the relationship between a pair of AUs can be complex and unique, existing approaches fail to specifically and explicitly represent such cues for each pair of AUs in each facial display. This paper proposes an AU relationship modelling approach that deep learns a unique graph to explicitly describe the relationship between each pair of AUs of the target facial display. Our approach first encodes each AU’s activation status and its association with other AUs into a node feature. Then, it learns a pair of multi-dimensional edge features to describe multiple task-specific relationship cues between each pair of AUs. During both node and edge feature learning, our approach also considers the influence of the unique facial display on AUs’ relationship by taking the full face representation as an input. Experimental results on BP4D and DISFA datasets show that both node and edge feature learning modules provide large performance improvements for CNN and transformer-based backbones, with our best systems achieving the state-of-the-art AU recognition results. Our approach not only has a strong capability in modelling relationship cues for AU recognition but also can be easily incorporated into various backbones. Our PyTorch code is made available at https://***/CVI-SZU/ME-GraphAU. Copyright © 2022, The Authors. All rights reserved.

关键词： Chemical activation

Towards AI-Powered Edge Intelligence for Object Detection in Self-Driving Cars: Enhancing IoV Efficiency and Safety

学校读者我要写书评

暂无评论

IEEE Internet of Things Journal 2025年

作者： Ahmed, Imran Ahmad, Misbah Siddiqi, Muftooh Ur Rehman Chehri, Abdellah Jeon, Gwangill Anglia Ruskin University School of Computing and Information Science Cambridge United Kingdom University of West of England Centre for Machine Vision Bristol Robotics Laboratory Bristol United Kingdom Hartpury University Department of Animal and Agriculture Gloucester United Kingdom Aston University College of Engineering and Physical Sciences Birmingham United Kingdom Department of Mathematics and Computer Science Canada Incheon National University Department of Embedded Systems Engineering Incheon Korea Republic of

In the rapidly advancing field of intelligent transportation systems, integrating Artificial Intelligence (AI) with edge computing presents a promising way to enhance the safety and efficiency of the Internet of Vehicles (IoV). This study explores and presents a deep learning based object detection model within an edge computing framework which aims to facilitate real time object detection in self driving cars. Using an urban traffic scenarios based dataset, our research shows the ability of the model to accurately detect and classify various objects important for autonomous driving. The YOLOv8 model is used in this work due to its optimal balance between accuracy and computational efficiency. This model has also demonstrated its worth by achieving good performance results, including an average precision of 0.79, a recall of 0.62, and an F1-score of 0.69. The results are demonstrated by a detailed confusion matrix, highlighting the model's effectiveness in complex driving environments and underscoring its reliability for in-vehicle deployment. By implementing AI directly on edge devices within vehicles, our approach might be helpful in significantly reducing latency, boosting decision-making speed, and enhancing data privacy by minimising dependence on cloud processing. The findings not only support the model's capabilities but also illustrate the practical benefits of edge intelligence in autonomous vehicles. These benefits, such as faster decision-making and improved data privacy, contribute effectively to the IoV infrastructure. This study marks a substantial step towards recognising the possibility of AI-enhanced edge computing in driving the next generation of autonomous vehicle technology. © 2025 IEEE.

关键词： Artificial Intelligence Autonomous Vehicles Deep leaning Edge Computing Internet of Vehicles