LocateAnything: Fast and High-Quality Vision-Language Grounding with Parallel Box Decoding

2026-05-26

AI Summary

This work addresses the limitations of conventional vision-language models in visual grounding and detection, which decode bounding boxes token-by-token in a sequential manner, thereby compromising geometric consistency and inference efficiency. To overcome these issues, the authors propose LocateAnything, a unified generative framework that introduces, for the first time, a Parallel Box Decoding (PBD) mechanism. This approach treats bounding boxes and other geometric primitives as atomic units, generating them in a single step to substantially enhance both geometric coherence and decoding parallelism. Coupled with a large-scale, high-quality vision-language data engine, the method achieves state-of-the-art performance across multiple benchmarks, simultaneously delivering high IoU-based localization accuracy and high throughput—effectively breaking the longstanding trade-off between speed and precision.

Abstract

Vision-language models (VLMs) commonly formulate visual grounding and detection as a coordinate-token generation problem, serializing each 2D box into multiple 1D tokens that are learned and decoded largely independently. This token-by-token decoding mismatches the coupled structure of box geometry and creates a practical inference bottleneck due to strictly sequential generation. We introduce LocateAnything, a unified generative grounding and detection framework based on Parallel Box Decoding (PBD). By decoding geometric elements such as bounding boxes and points as atomic units in a single step, LocateAnything preserves intra-box geometric coherence and unlocks substantial parallelism. We show that PBD improves both decoding throughput and localization accuracy. We further develop a scalable data engine and curate LocateAnything-Data, a large-scale dataset with more than 138 million training samples, substantially increasing data diversity for high-precision localization. Extensive evaluations show that LocateAnything advances the speed-accuracy frontier, achieving significantly higher decoding throughput while improving high-IoU localization quality across diverse benchmarks. The results highlight the complementary benefits of Parallel Box Decoding and large-scale training data in enabling efficient and precise unified visual grounding and detection.

Problem

• vision-language grounding
• bounding box decoding
• sequential generation
• geometric coherence
• inference bottleneck

Innovation

• Parallel Box Decoding
• Vision-Language Grounding
• Generative Detection
• High-Precision Localization
• Large-Scale Data Engine