Dynamic loss balancing and sequential enhancement for road-safety assessment and traffic scene classification

Fuente: arXiv
Saved in:
Bibliographic Details
Main Authors: Kačan, Marin, Ševrović, Marko, Šegvić, Siniša
Format: Preprint
Published: 2022
Subjects:
Online Access:
Tags: Add Tag
No Tags, Be the first to tag this record!
_version_ 1866929505098006528
author Kačan, Marin
Ševrović, Marko
Šegvić, Siniša
author_facet Kačan, Marin
Ševrović, Marko
Šegvić, Siniša
contents Road-safety inspection is an indispensable instrument for reducing road-accident fatalities contributed to road infrastructure. Recent work formalizes road-safety assessment in terms of carefully selected risk factors that are also known as road-safety attributes. In current practice, these attributes are manually annotated in geo-referenced monocular video for each road segment. We propose to reduce dependency on tedious human labor by automating recognition with a two-stage neural architecture. The first stage predicts more than forty road-safety attributes by observing a local spatio-temporal context. Our design leverages an efficient convolutional pipeline, which benefits from pre-training on semantic segmentation of street scenes. The second stage enhances predictions through sequential integration across a larger temporal window. Our design leverages per-attribute instances of a lightweight bidirectional LSTM architecture. Both stages alleviate extreme class imbalance by incorporating a multi-task variant of recall-based dynamic loss weighting. We perform experiments on the iRAP-BH dataset, which involves fully labeled geo-referenced video along 2,300 km of public roads in Bosnia and Herzegovina. We also validate our approach by comparing it with the related work on two road-scene classification datasets from the literature: Honda Scenes and FM3m. Experimental evaluation confirms the value of our contributions on all three datasets.
format Preprint
id arxiv_https___arxiv_org_abs_2211_04165
institution arXiv
publishDate 2022
record_format arxiv
spellingShingle Dynamic loss balancing and sequential enhancement for road-safety assessment and traffic scene classification
Kačan, Marin
Ševrović, Marko
Šegvić, Siniša
Computer Vision and Pattern Recognition
Road-safety inspection is an indispensable instrument for reducing road-accident fatalities contributed to road infrastructure. Recent work formalizes road-safety assessment in terms of carefully selected risk factors that are also known as road-safety attributes. In current practice, these attributes are manually annotated in geo-referenced monocular video for each road segment. We propose to reduce dependency on tedious human labor by automating recognition with a two-stage neural architecture. The first stage predicts more than forty road-safety attributes by observing a local spatio-temporal context. Our design leverages an efficient convolutional pipeline, which benefits from pre-training on semantic segmentation of street scenes. The second stage enhances predictions through sequential integration across a larger temporal window. Our design leverages per-attribute instances of a lightweight bidirectional LSTM architecture. Both stages alleviate extreme class imbalance by incorporating a multi-task variant of recall-based dynamic loss weighting. We perform experiments on the iRAP-BH dataset, which involves fully labeled geo-referenced video along 2,300 km of public roads in Bosnia and Herzegovina. We also validate our approach by comparing it with the related work on two road-scene classification datasets from the literature: Honda Scenes and FM3m. Experimental evaluation confirms the value of our contributions on all three datasets.
title Dynamic loss balancing and sequential enhancement for road-safety assessment and traffic scene classification
topic Computer Vision and Pattern Recognition
url https://arxiv.org/abs/2211.04165