正视差距：迁移标签以对齐目标检测数据集
Mind the Gap: Transferring Labels to Align Object Detection Datasets
对需要合并多源检测数据集、但目标标签空间固定的视觉研究者，LAT 提供了一条无需共享类别体系与人工重标注的标签迁移路径：用数据集专属检测器生成伪标签，再以特权提议生成器（PPG）替代 RPN、以语义特征融合（SFF）抑制伪标签噪声，可直接复用到跨域检测与数据扩充流程中。
原文依据 a Privileged Proposal Generator (PPG) that replaces the region proposal network in two-stage detectors
摘要译文
由模型翻译 · 术语统一 · 请对照原文核对关键结论融合多个目标检测数据集是提升模型泛化能力的一条途径，但类别语义与边界框标注的不一致阻碍了这一做法。部分应对方法假定各数据集共享标签体系，因而只处理空间层面的不一致；另一些方法需要人工重新标注，或生成一个统一标签空间，而在需要固定目标标签空间的场景下这可能并不适用。我们提出标签对齐迁移（Label-Aligned Transfer，LAT），一个标签迁移框架，系统性地将来自多个源数据集的标注投影到目标数据集的标签空间中。LAT 首先训练各数据集专属的检测器以生成伪标签，随后通过特权提议生成器（Privileged Proposal Generator，PPG）将这些伪标签与真实标注相结合，该生成器替代了两阶段检测器中的区域提议网络。为进一步精炼区域特征并应对伪标签噪声，语义特征融合（Semantic Feature Fusion，SFF）模块借助置信度加权的注意力机制，从相互重叠的提议中注入类别感知的上下文与特征。该流程在实现跨异构数据集的多对一标签空间迁移的同时，保留了各数据集特有的标注粒度，从而得到在语义与空间上均对齐、适合训练下游检测器的表示。因此，LAT 在不依赖共享标签空间、也不依赖人工重新标注的前提下，同时解决了类别层面的错位与边界框的不一致。在多个基准上，LAT 均带来检测性能的稳定提升，相较基线方法最高取得 +8.4 AP 的增益。
英文原文摘要 1615 字符
Combining multiple object detection datasets offers a path to improved model generalisation but is hindered by inconsistencies in class semantics and bounding box annotations. Some methods to address this assume shared label taxonomies and address only spatial inconsistencies; others require manual relabelling, or produce a unified label space, which may be unsuitable when a fixed target label space is required. We propose Label-Aligned Transfer (LAT), a label transfer framework that systematically projects annotations from diverse source datasets into the label space of a target dataset. LAT begins by training dataset-specific detectors to generate pseudo-labels, which are then combined with ground-truth annotations via a Privileged Proposal Generator (PPG) that replaces the region proposal network in two-stage detectors. To further refine region features and address pseudo-label noise, a Semantic Feature Fusion (SFF) module injects class-aware context and features from overlapping proposals using a confidence-weighted attention mechanism. This pipeline preserves dataset-specific annotation granularity while enabling many-to-one label space transfer across heterogeneous datasets, resulting in a semantically and spatially aligned representation suitable for training a downstream detector. LAT thus jointly addresses both class-level misalignments and bounding box inconsistencies without relying on shared label spaces or manual re-annotation. Across multiple benchmarks, LAT demonstrates consistent improvements in detection performance, achieving gains of up to +8.4 AP over baseline methods.