用深度特征的秩衡量图像视觉丰富度，无需标签且计算高效。
ERank in Latent Space as an Image-Complexity and Richness Measure
- 通过冻结编码器单次前向传播计算通道协方差的有效秩
- 与人类对复杂度的判断相关性达0.72，能区分图像丰富程度
- 适合图像超分和OCR任务中的数据筛选，提升模型性能
我们提出一种有效秩（ERank），作为图像深度特征图通道协方差的每样本、无标签视觉丰富度度量，仅需一次前向传播即可计算。ERank反映图像激活了多少个正交通道方向，我们分析其在噪声下的行为特性。实验表明，ERank能按从简单到丰富的顺序排列图像，与编码码率、清晰度、边缘密度相关，并与IC9600数据集上的人类复杂度标注有0.72的相关性。作为数据选择标准，剔除低ERank样本可提升超分辨率性能，剔除高ERank样本则有助于改善OCR表现，而在分类、分割和去噪任务中则无明显帮助。因此，ERank是一种低成本的丰富度信号，尤其适用于输入丰富度影响任务难度的场景。
原文摘要 · Abstract (English)
We propose the effective rank (ERank) of the channel covariance of an image's deep feature map as a per-sample, label-free measure of visual richness, computed from a single forward pass through a frozen pretrained encoder. ERank counts how many decorrelated channel directions an image activates, and we characterize its properties, including its behavior under noise. Empirically, ERank orders images from plain to visually rich, correlates with codec bitrate, sharpness, and edge density, and correlates with human complexity annotations on IC9600 with $r = 0.72$. As a data-selection criterion, removing low-ERank samples improves super-resolution and removing high-ERank samples improves OCR, in both pretraining and finetuning, while selection does not help classification, segmentation, or denoising. ERank is thus a cheap richness signal, useful exactly when task difficulty is governed by input richness.
Thank you to arXiv for use of its open access interoperability. PaperDance 不是 arXiv 官方产品；中文卡片由大模型生成，请以原文为准。