Multi-Channel Attention Selection GAN with Cascaded Semantic Guidance for Cross-View Image Translation

TL;DR

This paper introduces SelectionGAN, a two-stage generative model that leverages semantic maps and multi-channel attention to improve cross-view image translation, producing more accurate and realistic scene images from different viewpoints.

Contribution

The paper proposes a novel two-stage SelectionGAN with multi-channel attention and semantic guidance for enhanced cross-view image translation.

Findings

Outperforms state-of-the-art methods on multiple datasets

Generates more accurate and realistic images from different viewpoints

Utilizes uncertainty-guided pixel loss for better optimization

Abstract

Cross-view image translation is challenging because it involves images with drastically different views and severe deformation. In this paper, we propose a novel approach named Multi-Channel Attention SelectionGAN (SelectionGAN) that makes it possible to generate images of natural scenes in arbitrary viewpoints, based on an image of the scene and a novel semantic map. The proposed SelectionGAN explicitly utilizes the semantic information and consists of two stages. In the first stage, the condition image and the target semantic map are fed into a cycled semantic-guided generation network to produce initial coarse results. In the second stage, we refine the initial results by using a multi-channel attention selection mechanism. Moreover, uncertainty maps automatically learned from attentions are used to guide the pixel loss for better network optimization. Extensive experiments on Dayton, CVUSA and Ego2Top datasets show that our model is able to generate significantly better results than the state-of-the-art methods. The source code, data and trained models are available at https://github.com/Ha0Tang/SelectionGAN.