2023
Silva, A. L.; Oliveira, P.; Durães, D.; Fernandes, D.; Névoa, R.; Monteiro, J.; Melo-Pinto, P.; Machado, J.; Novais, P.
Em: Sensors, vol. 23, não 14, 2023, ISSN: 14248220, (cited By 1).
Resumo | Links | BibTeX | Etiquetas: 3D object; 3d object detection; Autonomous driving; Deep learning method; Learning methods; LiDAR sensing technology; Objects detection; Sensing technology; Software Specification; State of the art, article; deep learning; software, Autonomous vehicles; Deep learning; Learning systems; Object recognition; Optical radar; Specifications; Three dimensional computer graphics, Object detection
@article{Silva2023,
title = {A Framework for Representing, Building and Reusing Novel State-of-the-Art Three-Dimensional Object Detection Models in Point Clouds Targeting Self-Driving Applications},
author = {A. L. Silva and P. Oliveira and D. Durães and D. Fernandes and R. Névoa and J. Monteiro and P. Melo-Pinto and J. Machado and P. Novais},
url = {https://www.scopus.com/inward/record.uri?eid=2-s2.0-85166008165&doi=10.3390%2fs23146427&partnerID=40&md5=53b52635577fee2589e3dd1eca23b062},
doi = {10.3390/s23146427},
issn = {14248220},
year = {2023},
date = {2023-01-01},
journal = {Sensors},
volume = {23},
number = {14},
publisher = {Multidisciplinary Digital Publishing Institute (MDPI)},
abstract = {The rapid development of deep learning has brought novel methodologies for 3D object detection using LiDAR sensing technology. These improvements in precision and inference speed performances lead to notable high performance and real-time inference, which is especially important for self-driving purposes. However, the developments carried by these approaches overwhelm the research process in this area since new methods, technologies and software versions lead to different project necessities, specifications and requirements. Moreover, the improvements brought by the new methods may be due to improvements in newer versions of deep learning frameworks and not just the novelty and innovation of the model architecture. Thus, it has become crucial to create a framework with the same software versions, specifications and requirements that accommodate all these methodologies and allow for the easy introduction of new methods and models. A framework is proposed that abstracts the implementation, reusing and building of novel methods and models. The main idea is to facilitate the representation of state-of-the-art (SoA) approaches and simultaneously encourage the implementation of new approaches by reusing, improving and innovating modules in the proposed framework, which has the same software specifications to allow for a fair comparison. This makes it possible to determine if the key innovation approach outperforms the current SoA by comparing models in a framework with the same software specifications and requirements. © 2023 by the authors.},
note = {cited By 1},
keywords = {3D object; 3d object detection; Autonomous driving; Deep learning method; Learning methods; LiDAR sensing technology; Objects detection; Sensing technology; Software Specification; State of the art, article; deep learning; software, Autonomous vehicles; Deep learning; Learning systems; Object recognition; Optical radar; Specifications; Three dimensional computer graphics, Object detection},
pubstate = {published},
tppubtype = {article}
}
The rapid development of deep learning has brought novel methodologies for 3D object detection using LiDAR sensing technology. These improvements in precision and inference speed performances lead to notable high performance and real-time inference, which is especially important for self-driving purposes. However, the developments carried by these approaches overwhelm the research process in this area since new methods, technologies and software versions lead to different project necessities, specifications and requirements. Moreover, the improvements brought by the new methods may be due to improvements in newer versions of deep learning frameworks and not just the novelty and innovation of the model architecture. Thus, it has become crucial to create a framework with the same software versions, specifications and requirements that accommodate all these methodologies and allow for the easy introduction of new methods and models. A framework is proposed that abstracts the implementation, reusing and building of novel methods and models. The main idea is to facilitate the representation of state-of-the-art (SoA) approaches and simultaneously encourage the implementation of new approaches by reusing, improving and innovating modules in the proposed framework, which has the same software specifications to allow for a fair comparison. This makes it possible to determine if the key innovation approach outperforms the current SoA by comparing models in a framework with the same software specifications and requirements. © 2023 by the authors.
2022
Pereira, P.; Silva, A. Linhares; Machado, R.; Silva, J.; Durães, D.; Machado, J.; Novais, P.; Monteiro, J.; Melo-Pinto, P.; Fernandes, D.
Comparison of Different Deployment Approaches of FPGA-Based Hardware Accelerator for 3D Object Detection Models Proceedings Article
Em: G., Paiva A. Martins B. Marreiros (Ed.): pp. 285-296, Springer Science and Business Media Deutschland GmbH, 2022, ISSN: 03029743, (cited By 2; Conference of 21st EPIA Conference on Artificial Intelligence, EPIA 2022 ; Conference Date: 31 August 2022 Through 2 September 2022; Conference Code:283109).
Resumo | Links | BibTeX | Etiquetas: Autonomous vehicles; Computation theory; Computational efficiency; Convolution; Deep learning; Energy efficiency; Field programmable gate arrays (FPGA); Graphics processing unit; Integrated circuit design; Object recognition; Optical radar; Program processors, Autonomous Vehicles; Design and implementations; Detection models; FPGA-based hardware accelerators; Hardware accelerators; Hardware IP; Light detection and ranging; Objects detection; Real-time inference, Object detection
@inproceedings{Pereira2022285,
title = {Comparison of Different Deployment Approaches of FPGA-Based Hardware Accelerator for 3D Object Detection Models},
author = {P. Pereira and A. Linhares Silva and R. Machado and J. Silva and D. Durães and J. Machado and P. Novais and J. Monteiro and P. Melo-Pinto and D. Fernandes},
editor = {Paiva A. Martins B. Marreiros G.},
url = {https://www.scopus.com/inward/record.uri?eid=2-s2.0-85138741507&doi=10.1007%2f978-3-031-16474-3_24&partnerID=40&md5=cee8ca98871acfc2bf58e53a13e9eefa},
doi = {10.1007/978-3-031-16474-3_24},
issn = {03029743},
year = {2022},
date = {2022-01-01},
journal = {Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)},
volume = {13566 LNAI},
pages = {285-296},
publisher = {Springer Science and Business Media Deutschland GmbH},
abstract = {GPU servers have been responsible for the recent improvements in the accuracy and inference speed of the object detection models targeted to autonomous driving. However, its features, namely, power consumption and dimension, make its integration in autonomous vehicles impractical. Hybrid FPGA-CPU boards emerged as an alternative to server GPUs in the role of edge devices in autonomous vehicles. Despite their energy efficiency, such devices do not offer the same computational power as GPU servers and have fewer resources available. This paper investigates how to deploy deep learning models tailored to object detection in point clouds in edge devices for onboard real-time inference. Different approaches, requiring different levels of expertise in logic programming applied to FPGAs, are explored, resulting in three main solutions: utilization of software tools for model adaptation and compilation for a proprietary hardware IP; design and implementation of a hardware IP optimized for computing traditional convolutions operations; design and implementation of a hardware IP optimized for sparse convolutions operations. The performance of these solutions is compared in the KITTI dataset with computer performances. All the solutions resort to parallelism, quantization and optimized access control to memory to reduce the usage of logical FPGA resources, and improve processing time without significantly sacrificing accuracy. Solutions probed to be effective for real-time inference, power limited and space-constrained purposes. © 2022, The Author(s), under exclusive license to Springer Nature Switzerland AG.},
note = {cited By 2; Conference of 21st EPIA Conference on Artificial Intelligence, EPIA 2022 ; Conference Date: 31 August 2022 Through 2 September 2022; Conference Code:283109},
keywords = {Autonomous vehicles; Computation theory; Computational efficiency; Convolution; Deep learning; Energy efficiency; Field programmable gate arrays (FPGA); Graphics processing unit; Integrated circuit design; Object recognition; Optical radar; Program processors, Autonomous Vehicles; Design and implementations; Detection models; FPGA-based hardware accelerators; Hardware accelerators; Hardware IP; Light detection and ranging; Objects detection; Real-time inference, Object detection},
pubstate = {published},
tppubtype = {inproceedings}
}
GPU servers have been responsible for the recent improvements in the accuracy and inference speed of the object detection models targeted to autonomous driving. However, its features, namely, power consumption and dimension, make its integration in autonomous vehicles impractical. Hybrid FPGA-CPU boards emerged as an alternative to server GPUs in the role of edge devices in autonomous vehicles. Despite their energy efficiency, such devices do not offer the same computational power as GPU servers and have fewer resources available. This paper investigates how to deploy deep learning models tailored to object detection in point clouds in edge devices for onboard real-time inference. Different approaches, requiring different levels of expertise in logic programming applied to FPGAs, are explored, resulting in three main solutions: utilization of software tools for model adaptation and compilation for a proprietary hardware IP; design and implementation of a hardware IP optimized for computing traditional convolutions operations; design and implementation of a hardware IP optimized for sparse convolutions operations. The performance of these solutions is compared in the KITTI dataset with computer performances. All the solutions resort to parallelism, quantization and optimized access control to memory to reduce the usage of logical FPGA resources, and improve processing time without significantly sacrificing accuracy. Solutions probed to be effective for real-time inference, power limited and space-constrained purposes. © 2022, The Author(s), under exclusive license to Springer Nature Switzerland AG.