回到绘图板：重新思考场景级基于草图的图像检索
Back To The Drawing Board: Rethinking Scene-Level Sketch-Based Image Retrieval
AI总结：
针对场景级基于草图的图像检索，提出鲁棒训练目标，结合预训练、编码器与损失设计，在FS-COCO和SketchyCOCO上达到最先进性能，强调训练设计的关键作用。
摘要：
场景级基于草图的图像检索的目标是检索与手绘草图整体语义和空间布局相匹配的自然图像。与先前专注于检索模型架构增强的工作不同，我们强调真实世界草图中存在的固有模糊性和噪声。这一见解激发了一个明确设计为对草图变异性具有鲁棒性的训练目标。我们表明，通过预训练、编码器架构和损失函数的适当组合，可以在不引入额外复杂性的情况下实现最先进的性能。在具有挑战性的FS-COCO和广泛使用的SketchyCOCO数据集上进行的大量实验证实了我们方法的有效性，并强调了训练设计在跨模态检索任务中的关键作用，以及改进场景级SBIR评估场景的必要性。
英文摘要：
The goal of Scene-level Sketch-Based Image Retrieval is to retrieve natural images matching the overall semantics and spatial layout of a free-hand sketch. Unlike prior work focused on architectural augmentations of retrieval models, we emphasize the inherent ambiguity and noise present in real-world sketches. This insight motivates a training objective that is explicitly designed to be robust to sketch variability. We show that with an appropriate combination of pre-training, encoder architecture, and loss formulation, it is possible to achieve state-of-the-art performance without the introduction of additional complexity. Extensive experiments on a challenging FS-COCO and widely-used SketchyCOCO datasets confirm the effectiveness of our approach and underline the critical role of training design in cross-modal retrieval tasks, as well as the need to improve the evaluation scenarios of scene-level SBIR.
机构、总结、中文摘要均由AI分析而来，可能存在错误，请以论文正文为准。
