Mamba-YOLO-World: Marrying YOLO-World with Mamba for Open-Vocabulary Detection

TL;DR

Mamba-YOLO-World introduces a novel feature fusion mechanism that enhances open-vocabulary object detection performance while maintaining efficiency, outperforming previous models on COCO and LVIS benchmarks.

Contribution

The paper proposes MambaFusion-PAN, a new feature fusion architecture with linear complexity, improving YOLO-based open-vocabulary detection.

Findings

Outperforms YOLO-World on COCO and LVIS in zero-shot and fine-tuning.

Achieves better accuracy with fewer parameters and FLOPs.

Surpasses existing state-of-the-art OVD methods.

Abstract

Open-vocabulary detection (OVD) aims to detect objects beyond a predefined set of categories. As a pioneering model incorporating the YOLO series into OVD, YOLO-World is well-suited for scenarios prioritizing speed and efficiency. However, its performance is hindered by its neck feature fusion mechanism, which causes the quadratic complexity and the limited guided receptive fields. To address these limitations, we present Mamba-YOLO-World, a novel YOLO-based OVD model employing the proposed MambaFusion Path Aggregation Network (MambaFusion-PAN) as its neck architecture. Specifically, we introduce an innovative State Space Model-based feature fusion mechanism consisting of a Parallel-Guided Selective Scan algorithm and a Serial-Guided Selective Scan algorithm with linear complexity and globally guided receptive fields. It leverages multi-modal input sequences and mamba hidden states to guide the selective scanning process. Experiments demonstrate that our model outperforms the original YOLO-World on the COCO and LVIS benchmarks in both zero-shot and fine-tuning settings while maintaining comparable parameters and FLOPs. Additionally, it surpasses existing state-of-the-art OVD methods with fewer parameters and FLOPs.