论文标题
识别和减轻深度知觉相似性指标的缺陷
Identifying and Mitigating Flaws of Deep Perceptual Similarity Metrics
论文作者
论文摘要
测量图像的相似性是计算机视觉的一个基本问题，而计算机视觉不存在通用解决方案。尽管已显示出像素的简单指标，例如L2-NORM，但它们仍然存在很大的缺陷，但它们仍然受欢迎。一组最新的最新指标减轻了其中一些缺陷是深度知觉相似性（DPS）指标，其中相似性被评估为神经网络深度特征的距离。但是，DPS指标本身还没有彻底检查其利益，尤其是其缺陷。这项工作调查了最常见的DPS度量，其中通过空间位置比较了深层特征，并比较了平均和排序的深度特征。通过使用专门设计用于挑战它们的图像，对指标进行了深入分析，以了解指标的优势和劣势。这项工作有助于对DPS的缺陷进行新的见解，并进一步提出了对指标的改进。这项工作的实施可在线获得：https：//github.com/guspih/deep_perceptual_similarity_analysis/
Measuring the similarity of images is a fundamental problem to computer vision for which no universal solution exists. While simple metrics such as the pixel-wise L2-norm have been shown to have significant flaws, they remain popular. One group of recent state-of-the-art metrics that mitigates some of those flaws are Deep Perceptual Similarity (DPS) metrics, where the similarity is evaluated as the distance in the deep features of neural networks. However, DPS metrics themselves have been less thoroughly examined for their benefits and, especially, their flaws. This work investigates the most common DPS metric, where deep features are compared by spatial position, along with metrics comparing the averaged and sorted deep features. The metrics are analyzed in-depth to understand the strengths and weaknesses of the metrics by using images designed specifically to challenge them. This work contributes with new insights into the flaws of DPS, and further suggests improvements to the metrics. An implementation of this work is available online: https://github.com/guspih/deep_perceptual_similarity_analysis/
