Szczegóły publikacji
Opis bibliograficzny
LiFT: lightweight, FPGA-tailored 3D object detection based on LiDAR data / Konrad LIS, Tomasz KRYJAK, Marek GORGOŃ // W: Design and Architecture for Signal and Image Processing : 18th international workshop, DASIP 2025 : Barcelona, Spain, January 20–22, 2025 : proceedings / eds. Jordane Lorandel, Ahmed Kamaleldin. — Cham : Springer, cop. 2025. — ( Lecture Notes in Computer Science ; ISSN 0302-9743 ; 15569 ). — ISBN: 978-3-031-87896-1; e-ISBN: 978-3-031-87897-8. — S. 28–40. — Bibliogr., Abstr. — Publikacja dostępna online od: 2025-04-01
Autorzy (3)
Słowa kluczowe
Dane bibliometryczne
| ID BaDAP | 160941 |
|---|---|
| Data dodania do BaDAP | 2025-07-24 |
| DOI | 10.1007/978-3-031-87897-8_3 |
| Rok publikacji | 2025 |
| Typ publikacji | materiały konferencyjne (aut.) |
| Otwarty dostęp |  |
| Wydawca | Springer |
| Czasopismo/seria | Lecture Notes in Computer Science |
Abstract
This paper presents LiFT, a lightweight, fully quantized 3D object detection algorithm for LiDAR data, optimized for real-time inference on FPGA platforms. Through an in-depth analysis of FPGA-specific limitations, we identify a set of FPGA-induced constraints that shape the algorithm’s design. These include a computational complexity limit of 30 GMACs (billion multiply-accumulate operations), INT8 quantization for weights and activations, 2D cell-based processing instead of 3D voxels, and minimal use of skip connections. To meet these constraints while maximizing performance, LiFT combines novel mechanisms with state-of-the-art techniques such as reparameterizable convolutions and fully sparse architecture. Key innovations include the Dual-bound Pillar Feature Net, which boosts performance without increasing complexity, and an efficient scheme for INT8 quantization of input features. With a computational cost of just 20.73 GMACs, LiFT stands out as one of the few algorithms targeting minimal-complexity 3D object detection. Among comparable methods, LiFT ranks first, achieving an mAP of 51.84% and an NDS of 61.01% on the challenging NuScenes validation dataset. The code will be available at https://github.com/vision-agh/lift.
