English (United Kingdom)

https://curated-unify.zendy.io/wp-json/zendy-region/v1/featured_content/oa?rat=en

https://curated-unify.zendy.io/wp-json/zendy-region/v1/highlighted_journal/

Global: Zendy Plus annual

Presents the access of premium content as premium feature

Premium Content

Presents the keyphrase highlighting as premium feature

Keyphrase Highlighting

Presents the summarisation as premium feature

Summarisation

Insights

Presents the pdf analysis as premium feature

PDF Analysis

Presents the zaia usage as premium feature

ZAIA

Global: Zendy Plus monthly

Global: Zendy Tools annual

Global: Zendy Tools monthly

Global: Zendy Free Plan

Traditional video-based human activity recognition has experienced remarkableprogress linked to the rise of deep learning, but this effect was slower as itcomes to the downstream task of driver behavior understanding. Understandingthe situation inside the vehicle cabin is essential for Advanced DrivingAssistant System (ADAS) as it enables identifying distraction, predictingdriver's intent and leads to more convenient human-vehicle interaction. At thesame time, driver observation systems face substantial obstacles as they needto capture different granularities of driver states, while the complexity ofsuch secondary activities grows with the rising automation and increased driverfreedom. Furthermore, a model is rarely deployed under conditions identical tothe ones in the training set, as sensor placements and types vary from vehicleto vehicle, constituting a substantial obstacle for real-life deployment ofdata-driven models. In this work, we present a novel vision-based framework forrecognizing secondary driver behaviours based on visual transformers and anadditional augmented feature distribution calibration module. This moduleoperates in the latent feature-space enriching and diversifying the trainingset at feature-level in order to improve generalization to novel dataappearances, (e.g., sensor changes) and general feature quality. Our frameworkconsistently leads to better recognition rates, surpassing previousstate-of-the-art results of the public Drive&Act benchmark on all granularitylevels. Our code is publicly available athttps://github.com/KPeng9510/TransDARC.

TransDARC: Transformer-based Driver Activity Recognition with Latent Space Feature Calibration