LookAgain: Closed-Loop GUI Grounding with Visually Grounded Reflection
2026-08-10 • Computer Vision and Pattern Recognition
Computer Vision and Pattern Recognition
AI summaryⓘ
The authors identify a problem in current GUI grounding models where once a location is predicted, it cannot be reconsidered or corrected based on new visual information. To solve this, they propose LookAgain, a system that repeatedly predicts a position, checks it against the visual context, and refines the prediction if needed, creating a feedback loop. LookAgain uses two main steps: making a guess about a location and then confirming or rejecting it by looking again. Their experiments show that this approach improves accuracy, especially in challenging cases with small or crowded GUI elements.
GUI groundingpost-hoc reflectionvisual evidencecoordinate predictioniterative refinementreinforcement learningSFT (Supervised Fine-Tuning)GRPO (Goal-Reward Policy Optimization)closed-loop systemout-of-distribution
Authors
Renshan Zhang, Haoyang Meng, Yixiao He, Rui Shao, April Hua Liu, Liqiang Nie
Abstract
Recent graphical user interface (GUI) grounders have significantly advanced single-shot accuracy on standard benchmarks, yet their performance degrades sharply on small targets, densely packed controls and out-of-distribution interfaces. We attribute this gap to a paradigmatic limitation shared by existing approaches: none of them treats a produced coordinate as a hypothesis to be reflected upon and revised under new visual evidence. This manifests as three coupled issues: 1) Lack of post-hoc reflection. The prediction is frozen at the moment of emission, leaving no internal mechanism to challenge or refine it. 2) Visual evidence decoupled from the prediction. The auxiliary visual evidence is gathered to support the upcoming coordinate rather than to scrutinise the one already committed to. 3) Refinement over views, not over predictions. The iterative zoom-in refines the inspected region instead of inheriting a previous coordinate as a spatial prior to be corrected. In this paper, we propose LookAgain, a closed-loop GUI grounder driven by post-prediction visual reflection. LookAgain reformulates grounding as a multi-turn predict-look-again-refine process with two primitives: "locate" posts a coordinate hypothesis, renders a marker on the image and appends a local patch of the predicted region. It anchors the next reasoning step to the previous prediction as a spatial prior; "confirm" accepts or reject the hypothesis and terminates the procedure. We train the LookAgain grounder with SFT on constructed reflective trajectories as a cold start, followed by GRPO with terminal grounding correctness as the sole reward. Extensive experiments show that LookAgain consistently improves performance on both refusal-aware and general GUI grounding benchmarks, achieving state-of-the-art results. Comprehensive ablations further verify the effectiveness of the proposed framework.