n-Reference Transfer Learning for Saliency Prediction-猫眼课题宝

DOI：

10.1007/978-3-030-58598-3_30

发表时间：

2020-07

期刊：

ArXiv

影响因子：

通讯作者：

中科院分区：

其他

文献类型：

作者：

Yan Luo;Yongkang Wong;M. Kankanhalli;Qi Zhao

得益于深度学习研究和大规模数据集，显着性预测在过去十年中取得了巨大成功。然而，在缺乏足够数据用于数据饥渴模型的新领域中预测图像上的显着性图仍然具有挑战性。为了解决这个问题，我们提出了一种用于显着性预测的少量迁移学习范式，该范式能够将从现有的大规模显着性数据集学习到的知识有效地迁移到具有有限标记样本的目标域。具体地，使用很少的目标域样本作为参考来训练具有源域数据集的模型，使得训练过程可以收敛到有利于目标域的局部最小值。然后，学习模型进一步微调参考。所提出的框架是基于梯度和模型不可知的。我们对不同的源域和靶域对进行了全面的实验和烧蚀研究。结果表明，该框架实现了显著的性能提升。该代码可在以下网址公开获取： https://github.com/luoyan407/n-reference .

Benefiting from deep learning research and large-scale datasets, saliency prediction has achieved significant success in the past decade. However, it still remains challenging to predict saliency maps on images in new domains that lack sufficient data for data-hungry models. To solve this problem, we propose a few-shot transfer learning paradigm for saliency prediction, which enables efficient transfer of knowledge learned from the existing large-scale saliency datasets to a target domain with limited labeled samples. Specifically, few target domain samples are used as thereferenceto train a model with a source domain dataset such that the training process can converge to a local minimum in favor of the target domain. Then, the learned model is further fine-tuned with thereference. The proposed framework is gradient-based and model-agnostic. We conduct comprehensive experiments and ablation study on various source domain and target domain pairs. The results show that the proposed framework achieves a significant performance improvement. The code is publicly available at https://github.com/luoyan407/n-reference .

n-Reference Transfer Learning for Saliency Prediction

文献摘要

相似文献