# Detecting 32 Pedestrian Attributes for Autonomous Vehicles

arXiv: https://arxiv.org/abs/2012.02647 · HTML: https://mlsanity.com/paper/2012.02647/

## Key facts

- Cited by 10 papers in the map (data through 14 Oct 2026).
- First posted on arXiv 4 Dec 2020.

## Built on by (most cited first, 10 of 10)

- [Explainability of deep vision-based autonomous driving systems: Review and challenges](https://mlsanity.com/paper/2101.05307/) (2021)
- [FULLER: Unified Multi-modality Multi-task 3D Perception via Multi-level Gradient Calibration](https://mlsanity.com/paper/2307.16617/) (2023)
- [M3Net: Multimodal Multi-task Learning for 3D Detection, Segmentation, and Occupancy Prediction in Autonomous Driving](https://mlsanity.com/paper/2503.18100/) (2025)
- [Keypoint Communities](https://mlsanity.com/paper/2110.00988/) (2021)
- [Do Pedestrians Pay Attention? Eye Contact Detection in the Wild](https://mlsanity.com/paper/2112.04212/) (2021)
- [Pedestrian Stop and Go Forecasting with Hybrid Feature Fusion](https://mlsanity.com/paper/2203.02489/) (2022)
- [Local and Global Contextual Features Fusion for Pedestrian Intention Prediction](https://mlsanity.com/paper/2305.01111/) (2023)
- [Where Do You Go? Pedestrian Trajectory Prediction using Scene Features](https://mlsanity.com/paper/2501.13848/) (2025)
- [NITEC: Versatile Hand-Annotated Eye Contact Dataset for Ego-Vision Interaction](https://mlsanity.com/paper/2311.04505/) (2023)
- [Valeo Near-Field: a novel dataset for pedestrian intent detection](https://mlsanity.com/paper/2510.15673/) (2025)

## Abstract (arXiv)

Pedestrians are arguably one of the most safety-critical road users to consider for autonomous vehicles in urban areas. In this paper, we address the problem of jointly detecting pedestrians and recognizing 32 pedestrian attributes from a single image. These encompass visual appearance and behavior, and also include the forecasting of road crossing, which is a main safety concern. For this, we introduce a Multi-Task Learning (MTL) model relying on a composite field framework, which achieves both goals in an efficient way. Each field spatially locates pedestrian instances and aggregates attribute predictions over them. This formulation naturally leverages spatial context, making it well suited to low resolution scenarios such as autonomous driving. By increasing the number of attributes jointly learned, we highlight an issue related to the scales of gradients, which arises in MTL with numerous tasks. We solve it by normalizing the gradients coming from different objective functions when they join at the fork in the network architecture during the backward pass, referred to as fork-normalization. Experimental validation is performed on JAAD, a dataset providing numerous attributes for pedestrian analysis from autonomous vehicles, and shows competitive detection and attribute recognition results, as well as a more stable MTL training.