发表机构
Electronic Materials Research Laboratory, Key Laboratory of the Ministry of Education and International Center for Dielectric Research, School of Electronic Science and Engineering, Xi’an Jiaotong University; Micius Laboratory, Henan Academy of Sciences（电子材料研究中心、教育部重点实验室和国际介电研究中心、电子科学与工程学院，西安交通大学; 墨子实验室、河南省科学院）
机构由 AI 辅助整理，请以论文原文为准。AI 中文总结
研究无图像单像素传感中提升频谱对鲁棒性的影响，提出基于适应程度排序的方法，介绍时空软融合网络及训练方式，通过模拟和实际验证其性能，绘制提升频谱可指导设计，提升了该领域的方法效果。
AI 中文摘要
单像素传感将场景编码为短序列的编码测量值，无图像方法直接从该序列推断任务。去除重建并未消除困难，而是将其转移到提升过程，即将一维测量映射到二维表示的映射，先前工作将其视为简单的重塑。我们将提升过程重新塑造为无图像传感的核心设计轴，并根据其对输入的适应程度对方法进行排序：固定物理逆（先重建后分割）、学习到的静态投影或内容自适应检索；在此提升频谱上的位置可预测采集质量下降时的行为。时空软融合（STSF）网络在其U-Net++解码器之前，将探针选择的循环编码器与通过参数匹配消融选择的交叉注意力提升相结合，并在任务优先损失调度（TPLS，一种预定的重建先验）下进行训练。在模拟中，STSF+TPLS在三个数据集上以3.13%的采样率超过了先前的无图像基线（前景mIoU提高了3.2至9.9个百分点），并在降至0.39%时趋于平稳。最强的干净训练先重建后分割基线在无噪声极限下获胜，但在校准测量噪声下，无图像推理超过了它，原因是：重建管道在其分割器读取之前放大了相同的测量噪声。每个区域都以其独特的特征失败：崩溃、印记或粗糙化。STSF+TPLS作为概念验证，无需微调即可转移到真实的单像素平台，每个掩码约14毫秒。绘制提升频谱将分散的设计空间变成了一张在每个操作点应部署何种提升的地图。代码和预训练权重：此https URL
英文摘要
Single-pixel sensing encodes a scene as a short sequence of coded measurements, and image-free methods infer the task directly from that sequence. We show that removing image reconstruction relocates the central design problem to the lift: how 1D measurements become a 2D task representation. We organize this choice as a lift spectrum from a fixed-physics inverse, through a learned static projection, to content-adaptive retrieval. These are not interchangeable forms of reconstruction: the fixed-physics route reconstructs an image consumed at inference, whereas our spatiotemporal soft-fusion (STSF) network lifts measurements directly into task features, and task-prioritized loss scheduling (TPLS) uses a separate learned reconstruction branch only as scheduled training supervision. A probe-selected recurrent encoder and a parameter-matched lift ablation identify the STSF design. In simulation, STSF+TPLS exceeds the prior image-free baseline on three datasets at 3.13% sampling (+3.2 to +9.9 pp foreground mIoU) and remains competitive down to 0.39%. The strongest clean-trained reconstruct-then-segment baseline wins without measurement noise, but measurement noise reverses the ranking: the reconstructed task input carries a 20-70x larger normalized relative perturbation than the measurements themselves. Stressed to failure, the three lift regions exhibit distinct dominant signatures--collapse, imprinting, and coarsening. STSF+TPLS transfers without fine-tuning to a real single-pixel bench, where the reversal reappears as a proof of concept; inference takes about 14 ms per mask on an RTX 4090. Within the tested fixed-acquisition regime, measurement-to-space adaptivity therefore organizes both the clean-to-noisy operating envelope and the failure a system encounters. Code and pretrained weights: https://github.com/Hanyuyuan6/STSF-TPLS.
Comments25 pages (13 main text + 12 supplementary material), 8 figures, 3 tables. Submitted to IEEE Transactions on Computational Imaging
