PISA: A Pseudo-Individual Source-Domain Feature Adaptation Framework for Test-Time Open-Vocabulary Object Detection

📄 arXiv: 2608.14142v1 📥 PDF

作者: Ziyan He, Xiongtai Yang, Tao Wang

分类: cs.CV

发布日期: 2026-08-14


💡 一句话要点

提出PISA框架以解决开放词汇目标检测中的性能下降问题

🎯 匹配领域: 支柱三:空间感知与语义 (Perception & Semantics)

关键词: 开放词汇目标检测 测试时自适应 特征对齐 抗干扰特征提取 深度学习 计算机视觉 无源学习

📋 核心要点

  1. 现有的无源OVOD-TTA方法在初始预测不佳时依赖伪标签,导致准确性显著下降。
  2. PISA通过抗干扰特征提取器和特征对齐模块,将抗干扰特征转化为伪个体源域特征,增强检测性能。
  3. 在VOC-C、COCO-C和LVIS-C基准测试中,PISA显著提升了定位精度和类别识别准确性,表现优于现有方法。

📝 摘要(中文)

开放词汇目标检测测试时自适应(OVOD-TTA)旨在解决预训练基础模型在遇到图像域偏移时的性能下降。现有的无源OVOD-TTA方法依赖于精细的测试时信息或伪标签进行自训练,导致初始预测不佳时准确性显著下降。为此,本文提出PISA,一种新颖的无源OVOD-TTA方法,能够无缝集成到开放词汇视觉骨干网络中。核心组件包括抗干扰特征提取器(CIFE)、特征对齐模块(FAM)和多尺度对齐框架(BAA),通过这些组件,PISA显著提高了原始模型的定位精度和类别识别准确性,并在COCO-C数据集上超越现有方法3.92%的AP@50%表现。

🔬 方法详解

问题定义:本文旨在解决开放词汇目标检测中,预训练模型在图像域偏移时性能下降的问题。现有方法在初始预测不佳时,依赖伪标签进行自训练,导致准确性下降。

核心思路:PISA框架通过抗干扰特征提取器(CIFE)提取在不同干扰下保持不变的特征,结合特征对齐模块(FAM)和多尺度对齐框架(BAA),将这些特征转化为伪个体源域特征,以增强检测性能。

技术框架:PISA的整体架构包括三个主要模块:CIFE用于提取抗干扰特征,FAM用于特征对齐,BAA用于多尺度特征的整合与对齐。该框架能够在不依赖源域数据的情况下进行有效的特征适应。

关键创新:PISA的核心创新在于利用CLIP视觉特征的抗干扰性,确保在多种干扰下的鲁棒性,并通过伪个体源域特征替代不可靠的伪标签信号。

关键设计:在设计中,CIFE专注于提取密集且具体的特征,FAM和BAA则负责特征的对齐与转换,确保最终输出的特征能够有效支持目标检测任务。

🖼️ 关键图片

fig_0
fig_1
fig_2

📊 实验亮点

PISA在VOC-C、COCO-C和LVIS-C基准测试中表现出色,尤其在COCO-C数据集上,AP@50%提升了3.92%,显著超过了现有的无源OVOD-TTA方法,展示了其在目标检测领域的领先性能。

🎯 应用场景

PISA框架在开放词汇目标检测领域具有广泛的应用潜力,尤其适用于需要在不同环境下进行实时目标检测的场景,如自动驾驶、监控系统和机器人视觉等。该研究的成果能够提升模型在实际应用中的鲁棒性和准确性,推动智能视觉系统的发展。

📄 摘要(原文)

Open-vocabulary object detection test-time adaptation (OVOD-TTA) aims to address the performance degradation that pre-trained base models suffer when encountering image-domain shifts. Existing source-free OVOD-TTA methods rely either on refined test-time information for re-scoring or on pseudo-labels for self-training, leading to significant accuracy degradation when initial predictions are poor. Meanwhile, most conventional source-domain estimation methods recover abstract, sparse representations suitable for the classification task, but fail to capture the dense, concrete features required for detection. To address these issues, we propose PISA, a novel source-free OVOD-TTA method that can be seamlessly integrated into open-vocabulary visual backbones. The core components of our method are the Corruption-Invariant Feature Extractor (CIFE), the Feature Alignment Module (FAM), and a multi-scale alignment framework (BAA). To capture detection-suitable features, we develop CIFE to exploit the invariance of CLIP's visual features across corrupted images, ensuring robustness against various corruptions. We further develop FAM and BAA for the pre-training and adaptation to transform the corruption-invariant features into pseudo-individual source-domain features that are close to the original source-domain features. In this way, dense and concrete pseudo-individual source-domain features are used for supervision instead of unreliable pseudo-label signals. Experiments on the corrupted VOC-C, COCO-C, and LVIS-C benchmarks across three base models demonstrate that PISA substantially improves both the localization precision and the category recognition accuracy of the original models. Notably, PISA achieves state-of-the-art performance without requiring access to source-domain data, surpassing existing methods by 3.92% in AP@50% on COCO-C.