Search papers, labs, and topics across Lattice.
This paper introduces ReferTrack, a novel referring-then-tracking paradigm for embodied visual tracking (EVT) that utilizes a single forward-facing camera to identify and track targets described in natural language. By integrating a sliding-window queue of previously selected bounding boxes and employing temporal-viewpoint-bbox indicator (TVBI) tokens, ReferTrack effectively preserves motion cues and enhances target identification through co-training on a custom dataset. The model achieves state-of-the-art performance on EVT-Bench, with success rates surpassing multi-camera baselines, and demonstrates strong real-world applicability on legged and humanoid robots.
Achieving 89.4% success in single-view embodied visual tracking, ReferTrack outperforms multi-camera systems by grounding tracking decisions in real-time visual data.
Embodied visual tracking (EVT) requires a mobile agent to continuously follow a specific target described in natural language using only onboard vision. While recent vision-language-action (VLA) policies unify target identification and trajectory planning, their chain-of-thought (CoT) reasoning often operates in abstract spatial latents that are difficult to supervise and weakly aligned with explicit image-space detections. To address this, we introduce ReferTrack, a referring-then-tracking paradigm that grounds EVT using a single forward-facing camera. Our model first selects the target from an indexed set of bounding boxes, then decodes tracking waypoints conditioned on this image-grounded decision. To preserve target motion cues over time, ReferTrack maintains a sliding-window queue of previously selected bounding boxes, injecting their geometric features into the visual history via temporal-viewpoint-bbox indicator (TVBI) tokens. We further enhance target identification by co-training on a custom Refer-QA dataset. On EVT-Bench, ReferTrack achieves state-of-the-art single-view performance with success rates of 89.4%, 73.3%, and 74.1% on the single-target, distracted, and ambiguity tracking splits, respectively -- matching or even surpassing several multi-camera baselines on identification-heavy tasks. Finally, real-world deployments on legged and humanoid robots validate its robust sim-to-real transfer capabilities. Code is available at https://github.com/MedlarTea/referTrack.
