DINO-Explorer: Active Underwater Discovery via Ego-Motion Compensated Semantic Predictive Coding

arXiv cs.RO / 4/15/2026

📰 NewsSignals & Early TrendsIdeas & Deep AnalysisModels & Research

共有:

Key Points

DINO-Explorer is presented as an underwater active perception framework that prioritizes scientifically relevant, transient events instead of passively logging exhaustive video for offline review.
The method generates a continuous semantic “surprise” signal inside the latent space of a frozen DINOv3 foundation model, using a lightweight action-conditioned recurrent predictor to forecast short-horizon semantic changes.
An efference-copy-inspired module compensates for ego-motion by using globally pooled optical flow to suppress self-induced visual variation while preserving true environmental novelty.
Evaluations on asynchronous event triage under constrained telemetry show bandwidth-efficient attention, retaining 78.8% of consensus events with a 56.8% trigger confirmation rate.
Ego-motion conditioning reduces false positives by 45.5% versus an uncompensated surprise baseline, and a replay-side study reports 48.2% telemetry bandwidth reduction while maintaining a 62.2% peak F1 score near the selected operating point.

Abstract

Marine ecosystem degradation necessitates continuous, scientifically selective underwater monitoring. However, most autonomous underwater vehicles (AUVs) operate as passive data loggers, capturing exhaustive video for offline review and frequently missing transient events of high scientific value. Transitioning to active perception requires a causal, online signal that highlights significant phenomena while suppressing maneuver-induced visual changes. We propose DINO-Explorer, a novelty-aware perception framework driven by a continuous semantic surprise signal. Operating within the latent space of a frozen DINOv3 foundation model, it leverages a lightweight, action-conditioned recurrent predictor to anticipate short-horizon semantic evolution. An efference-copy-inspired module utilizes globally pooled optical flow to discount self-induced visual changes without suppressing genuine environmental novelty. We evaluate this signal on the downstream task of asynchronous event triage under variant telemetry constraints. Results demonstrate that DINO-Explorer provides a robust, bandwidth-efficient attention mechanism. At a fixed operating point, the system retains 78.8% of post-discovery human-reviewer consensus events with a 56.8% trigger confirmation rate, effectively surfacing mission-relevant phenomena. Crucially, ego-motion conditioning suppresses 45.5% of false positives relative to an uncompensated surprise signal baseline. In a replay-side Pareto ablation study, DINO-Explorer robustly dominates the validated peak F1 versus telemetry bandwidth frontier, reducing telemetry bandwidth by 48.2% at the selected operating point while maintaining a 62.2% peak F1 score, successfully concentrating data transmission around human-verified novelty events.