Exocentric to Egocentric Image Generation Via Parallel Generative Adversarial Network
Exocentric to Egocentric Image Generation Via Parallel Generative Adversarial Network
复制标题
DOI：
10.1109/icassp40776.2020.9053957
复制
发表时间：
2020-02
期刊：
影响因子：
--
通讯作者：
Gaowen Liu;Hao Tang;Hugo Latapie;Yan Yan-Yan
中科院分区：
文献类型：
--
作者：
Gaowen Liu;Hao Tang;Hugo Latapie;Yan Yan-Yan
最近提出了跨视图图像生成来从一个显着不同的视图生成一个视图的图像。在本文中，我们研究了外中心（第三人称）视图到自我中心（第一人称）视图图像的生成。这是一项具有挑战性的任务，因为自我中心的观点有时与外中心的观点截然不同。因此，在两个视图之间转换外观是一项艰巨的任务。为此，我们提出了一种新颖的并行生成对抗网络（P-GAN），具有新颖的跨周期损失，以学习从外心视图生成以自我为中心的图像的共享信息。我们还在学习过程中引入了一种新颖的上下文特征损失，以捕获图像中的上下文信息。对 Exo-Ego 数据集 [1] 的大量实验表明，我们的模型优于最先进的方法。
Cross-view image generation has been recently proposed to generate images of one view from another dramatically different view. In this paper, we investigate exocentric (third-person) view to egocentric (first-person) view image generation. This is a challenging task since egocentric view sometimes is remarkably different from exocentric view. Thus, transforming the appearances across the two views is a nontrivial task. To this end, we propose a novel Parallel Generative Adversarial Network (P-GAN) with a novel cross-cycle loss to learn the shared information for generating egocentric images from exocentric view. We also incorporate a novel contextual feature loss in the learning procedure to capture the contextual information in images. Extensive experiments on the Exo-Ego datasets [1] show that our model outperforms the state-of-the-art approaches.