arXivDaily arXiv每日学术速递 周一至周五更新
arXiv 2104.10064cs.CV
收藏

用于改进任意风格传输的风格感知归一化损失

Style-Aware Normalized Loss for Improving Arbitrary Style Transfer

USC Information Sciences Institute(南加州大学信息科学研究所);Amazon Alexa Natural Understanding(亚马逊Alexa自然理解部门)

Jiaxin Cheng, Ayush Jaiswal, Yue Wu, Pradeep Natarajan, Prem Natarajan

摘要:

神经风格迁移(NST)已经从单一风格迅速发展到无限风格模型,也称为任意风格迁移(AST)。虽然吸引人的结果已被广泛报道的文献中,我们的经验研究四个著名的AST方法(GoogleMagenta,AdaIN,LinearTransfer,和SANet)显示,超过50%的时间,AST风格化的图像是不可接受的人类用户,通常是由于下或过度风格化。我们系统地研究了这种不平衡的风格可移植性(IST)的原因,并提出了一个简单而有效的解决方案,以减轻这个问题。我们的研究表明,IST问题与传统的AST风格损失有关,并揭示了根本原因是训练样本的权重相等,而不管其对应的风格图像的属性如何,这使模型偏向于某些风格。通过研究AST型损失的理论界,我们提出了一种新的损失,在很大程度上克服了IST。理论分析和实验结果验证了我们的损失的有效性,超过80%的风格欺骗率和98%的相对较高的偏好,在人类的评价。

英文摘要:

Neural Style Transfer (NST) has quickly evolved from single-style to infinite-style models, also known as Arbitrary Style Transfer (AST). Although appealing results have been widely reported in literature, our empirical studies on four well-known AST approaches (GoogleMagenta, AdaIN, LinearTransfer, and SANet) show that more than 50% of the time, AST stylized images are not acceptable to human users, typically due to under- or over-stylization. We systematically study the cause of this imbalanced style transferability (IST) and propose a simple yet effective solution to mitigate this issue. Our studies show that the IST issue is related to the conventional AST style loss, and reveal that the root cause is the equal weightage of training samples irrespective of the properties of their corresponding style images, which biases the model towards certain styles. Through investigation of the theoretical bounds of the AST style loss, we propose a new loss that largely overcomes IST. Theoretical analysis and experimental results validate the effectiveness of our loss, with over 80% relative improvement in style deception rate and 98% relatively higher preference in human evaluation.

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。

↑