Infothek
This page contains automatically translated content.
New article at the "Workshop on roBustness and Reliability of Autonomous Vehicles in the Open-world (BRAVO), ICCV"
To address safety concerns in automated driving, a novel metric called IoUw is proposed in a new article at the "Workshop on roBustness and Reliability of Autonomous Vehicles in the Open-world (BRAVO), ICCV" from Jasmin Breitenstein, Florian Heidecker, Maria Lyssenko, Daniel Bogdoll, Maarten Bieshaar, J. Marius Zöllner, Bernhard Sick, and Tim Fingscheidt. This strategy focuses on the pixel-level relevance of semantic segmentation outputs. This metric, emphasizing small areas affected by corner cases, introduces a severity penalty during evaluation to accurately identify safety-critical misdetections and mitigate potential risks to vulnerable road users in large-scale datasets.
News
New article at the "Workshop on roBustness and Reliability of Autonomous Vehicles in the Open-world (BRAVO), ICCV"
To address safety concerns in automated driving, a novel metric called IoUw is proposed in a new article at the "Workshop on roBustness and Reliability of Autonomous Vehicles in the Open-world (BRAVO), ICCV" from Jasmin Breitenstein, Florian Heidecker, Maria Lyssenko, Daniel Bogdoll, Maarten Bieshaar, J. Marius Zöllner, Bernhard Sick, and Tim Fingscheidt. This strategy focuses on the pixel-level relevance of semantic segmentation outputs. This metric, emphasizing small areas affected by corner cases, introduces a severity penalty during evaluation to accurately identify safety-critical misdetections and mitigate potential risks to vulnerable road users in large-scale datasets.
Dates
New article at the "Workshop on roBustness and Reliability of Autonomous Vehicles in the Open-world (BRAVO), ICCV"
To address safety concerns in automated driving, a novel metric called IoUw is proposed in a new article at the "Workshop on roBustness and Reliability of Autonomous Vehicles in the Open-world (BRAVO), ICCV" from Jasmin Breitenstein, Florian Heidecker, Maria Lyssenko, Daniel Bogdoll, Maarten Bieshaar, J. Marius Zöllner, Bernhard Sick, and Tim Fingscheidt. This strategy focuses on the pixel-level relevance of semantic segmentation outputs. This metric, emphasizing small areas affected by corner cases, introduces a severity penalty during evaluation to accurately identify safety-critical misdetections and mitigate potential risks to vulnerable road users in large-scale datasets.