Search papers, labs, and topics across Lattice.
This paper introduces CoRE, a weakly supervised framework that enables fine-grained risk evidence learning from coarse video-level supervision in driving videos. By first training a video-level predictor and then utilizing structured interventions to assess the impact of specific temporal regions and entities, CoRE effectively distills this information into a model that predicts risk support without requiring detailed annotations at inference. The evaluation across multiple datasets demonstrates that CoRE achieves strong temporal localization and competitive performance, highlighting its ability to extract valuable insights from limited supervision.
CoRE reveals that coarse video predictions can effectively guide the identification of fine-grained risk evidence, eliminating the need for detailed annotations.
Perceived risk in driving evolves over time and may be supported by specific scene entities, yet supervision is typically limited to coarse video-level judgments. Learning \emph{when} supporting evidence emerges and \emph{which entities} support a risk predictor would ordinarily require costly temporal- and entity-level annotations. We introduce \textbf{CoRE}, a weakly supervised coarse-to-fine framework that learns fine-grained prediction support from coarse video supervision. CoRE first trains a video-level predictor and then freezes it. Structured interventions over candidate temporal regions or entity tracks measure how each candidate changes the coarse prediction, producing graded prediction-effect targets. These targets are distilled into a student that directly predicts temporal and entity support from the original video, without requiring interventions at inference. We evaluate this learning principle across three complementary settings: RISEE tests perceived-risk support from subjective clip-level judgments without temporal or entity-level risk annotations; DoTA provides independent temporal event annotations for evaluating weakly supervised traffic-anomaly localization; and UCF-Crime tests whether the same coarse-to-fine mechanism extends to a standard non-driving anomaly-detection benchmark. Across these settings, CoRE learns informative fine-grained support from coarse supervision, with strong temporal localization on DoTA and competitive performance on UCF-Crime. These results show that coarse video predictions can provide useful supervision for recovering the fine-grained evidence supporting them, without requiring corresponding fine-grained labels.
