Image Generation from Image Captioning -- Invertible Approach

TL;DR

A simple invertible neural network architecture is proposed for the dual tasks of image captioning and image generation while being trained on only one task.

Abstract

Our work aims to build a model that performs dual tasks of image captioning and image generation while being trained on only one task. The central idea is to train an invertible model that learns a one-to-one mapping between the image and text embeddings. Once the invertible model is efficiently trained on one task, the image captioning, the same model can generate new images for a given text through the inversion process, with no additional training. This paper proposes a simple invertible neural network architecture for this problem and presents our current findings.

Figures (2)

• Figure 1: Architecture for training invertible neural networks to predict caption embeddings
• Figure 2: Flow of inverse task: Generating an image from a given caption