TY - JOUR
T1 - FCOSH
T2 - A novel single-head FCOS for faster object detection in autonomous-driving systems
AU - Mboutayeb, Saad
AU - Majda, Aicha
AU - Zenkouar, Khalid
AU - Nikolov, Nikola S.
N1 - Publisher Copyright:
© 2023 The Author(s)
PY - 2024/3
Y1 - 2024/3
N2 - In autonomous driving systems, object detection plays a pivotal role by facilitating their ability to perceive the surrounding road environment effectively. Object detection's foremost challenge pertains to its real-time operational capabilities. Achieving this necessitates reducing the detectors' computational complexity while preserving their accuracy. Nevertheless, most of the approach in object detection involves dividing image processing over multiple heads, each tasked with detecting objects at particular scales. Even though this approach improves detection accuracy, it adds an extra computational burden. In this study, our objective is to assess the feasibility of employing a single head within the originally multi-headed architecture of the FCOS detector. In response to the challenges posed by this significant modification, we propose a set of straightforward solutions, resulting in the development of a novel Fully Convolutional One-Stage with a Single Head (FCOSH) detector. Through experiments on the BDD100K benchmark, our FCOSH detector exhibits substantial improvements in computational efficiency relative to the original FCOS while concurrently achieving a superior detection 0.5% accuracy. Specifically, FCOSH achieves an 18% reduction in inference time, a 24% reduction in required FLOPs, and a 10% decrease in the number of model parameters compared to FCOS.
AB - In autonomous driving systems, object detection plays a pivotal role by facilitating their ability to perceive the surrounding road environment effectively. Object detection's foremost challenge pertains to its real-time operational capabilities. Achieving this necessitates reducing the detectors' computational complexity while preserving their accuracy. Nevertheless, most of the approach in object detection involves dividing image processing over multiple heads, each tasked with detecting objects at particular scales. Even though this approach improves detection accuracy, it adds an extra computational burden. In this study, our objective is to assess the feasibility of employing a single head within the originally multi-headed architecture of the FCOS detector. In response to the challenges posed by this significant modification, we propose a set of straightforward solutions, resulting in the development of a novel Fully Convolutional One-Stage with a Single Head (FCOSH) detector. Through experiments on the BDD100K benchmark, our FCOSH detector exhibits substantial improvements in computational efficiency relative to the original FCOS while concurrently achieving a superior detection 0.5% accuracy. Specifically, FCOSH achieves an 18% reduction in inference time, a 24% reduction in required FLOPs, and a 10% decrease in the number of model parameters compared to FCOS.
KW - Anchor-free
KW - Autonomous-driving systems
KW - Fully convolutional
KW - Object detection
KW - One-stage
KW - Single-head
UR - http://www.scopus.com/inward/record.url?scp=85182022920&partnerID=8YFLogxK
U2 - 10.1016/j.iswa.2023.200324
DO - 10.1016/j.iswa.2023.200324
M3 - Article
AN - SCOPUS:85182022920
SN - 2667-3053
VL - 21
JO - Intelligent Systems with Applications
JF - Intelligent Systems with Applications
M1 - 200324
ER -