[Submitted on 8 Jun 2026 (v1), last revised 23 Oct 2026 (this version, v3)]
Abstract:Video retrieval at scale is central to data curation and safety validation in autonomous driving, where users want to find not only scenes but also dynamic events such as cut-ins and hard braking. Existing vision-language and keyword-based retrieval methods often miss these events because the relevant motion may not be explicitly described in text or captured by lexical overlap. Rule-based retrieval can encode such events more directly, but it is brittle: generated or hand-written rules often fail when their assumptions do not match real driving data. We propose STRIVE-D, a data-calibrated retrieval framework for driving videos. It uses weakly labeled in-domain videos to estimate when a query rule is reliable, adapt rules that mismatch observed data, and fuse calibrated rule scores with vision-language and keyword-based retrieval signals. Across three driving benchmarks, including newly released human-annotated event data on DrivingDojo, STRIVE-D delivers up to 84% relative improvement in top-1 accuracy over state-of-the-art methods.
Submission history
From: Manyi Yao [view email]
[v1] Mon, 8 Jun 2026 07:00:33 UTC (2,727 KB)
[v2] Sat, 26 Sep 2026 08:42:16 UTC (2,727 KB)
[v3] Fri, 23 Oct 2026 17:49:29 UTC (2,718 KB)
