Vehicle viewpoint estimation from monocular images is a crucial component for autonomous driving vehicles and for fleet management applications. In this paper, we make several contributions to advance the state-of-the-art on this problem. We show the effectiveness of applying a smoothing filter to the output neurons of a Convolutional Neural Network (CNN) when estimating vehicle viewpoint. We point out the overlooked fact that, under the same viewpoint, the appearance of a vehicle is strongly influenced by its position in the image plane, which renders viewpoint estimation from appearance an ill-posed problem. We show how, by inserting in the model a CoordConv layer to provide the coordinates of the vehicle, we are able to solve such ambiguity and greatly increase performance. Finally, we introduce a new data augmentation technique that improves viewpoint estimation on vehicles that are closer to the camera or partially occluded. All these improvements let a lightweight CNN reach optimal results while keeping inference time low. An extensive evaluation on a viewpoint estimation benchmark and on actual vehicle camera data shows that our method significantly outperforms the state-of-the-art in vehicle viewpoint estimation, both in terms of accuracy and memory footprint.

Magistri, S., Boschi, M., Sambo, F., de Andrade, D.C., Simoncini, M., Kubin, L., et al. (2022). Lightweight and Effective Convolutional Neural Networks for Vehicle Viewpoint Estimation From Monocular Images. IEEE TRANSACTIONS ON INTELLIGENT TRANSPORTATION SYSTEMS, 24(1), 191-200 [10.1109/TITS.2022.3216359].

Lightweight and Effective Convolutional Neural Networks for Vehicle Viewpoint Estimation From Monocular Images

Boschi, Marco
Co-primo
;
Luigi, Luca De
Penultimo
;
Salti, Samuele
Ultimo
2022

Abstract

Vehicle viewpoint estimation from monocular images is a crucial component for autonomous driving vehicles and for fleet management applications. In this paper, we make several contributions to advance the state-of-the-art on this problem. We show the effectiveness of applying a smoothing filter to the output neurons of a Convolutional Neural Network (CNN) when estimating vehicle viewpoint. We point out the overlooked fact that, under the same viewpoint, the appearance of a vehicle is strongly influenced by its position in the image plane, which renders viewpoint estimation from appearance an ill-posed problem. We show how, by inserting in the model a CoordConv layer to provide the coordinates of the vehicle, we are able to solve such ambiguity and greatly increase performance. Finally, we introduce a new data augmentation technique that improves viewpoint estimation on vehicles that are closer to the camera or partially occluded. All these improvements let a lightweight CNN reach optimal results while keeping inference time low. An extensive evaluation on a viewpoint estimation benchmark and on actual vehicle camera data shows that our method significantly outperforms the state-of-the-art in vehicle viewpoint estimation, both in terms of accuracy and memory footprint.
2022
Magistri, S., Boschi, M., Sambo, F., de Andrade, D.C., Simoncini, M., Kubin, L., et al. (2022). Lightweight and Effective Convolutional Neural Networks for Vehicle Viewpoint Estimation From Monocular Images. IEEE TRANSACTIONS ON INTELLIGENT TRANSPORTATION SYSTEMS, 24(1), 191-200 [10.1109/TITS.2022.3216359].
Magistri, Simone; Boschi, Marco; Sambo, Francesco; de Andrade, Douglas Coimbra; Simoncini, Matteo; Kubin, Luca; Taccari, Leonardo; Luigi, Luca De; Sal...espandi
File in questo prodotto:
File Dimensione Formato  
2022_T_ITS_Transaction__Viewpoint_Estimation_compressed.pdf

accesso aperto

Tipo: Postprint
Licenza: Licenza per accesso libero gratuito
Dimensione 3.93 MB
Formato Adobe PDF
3.93 MB Adobe PDF Visualizza/Apri

I documenti in IRIS sono protetti da copyright e tutti i diritti sono riservati, salvo diversa indicazione.

Utilizza questo identificativo per citare o creare un link a questo documento: https://hdl.handle.net/11585/905484
Citazioni
  • ???jsp.display-item.citation.pmc??? ND
  • Scopus 2
  • ???jsp.display-item.citation.isi??? 1
social impact