Large Hole Image Inpainting With Compress-Decompression Network

TL;DR

This paper introduces a compression-decompression neural network for large-hole image inpainting, effectively handling extensive missing regions by combining residual-based compression with texture-aware decompression, outperforming existing methods.

Contribution

The novel compression-decompression network improves large-hole inpainting by integrating residual compression and texture selection, surpassing traditional encoder-decoder models in handling extensive missing areas.

Findings

Better performance on Places2 and CelebA datasets.

More effective inpainting with many conflicts.

Outperforms existing methods in large-hole scenarios.

Abstract

Image inpainting technology can patch images with missing pixels. Existing methods propose convolutional neural networks to repair corrupted images. The networks focus on the valid pixels around the missing pixels, use the encoder-decoder structure to extract valuable information, and use the information to fix the vacancy. However, if the missing part is too large to provide useful information, the result will exist blur, color mixing, and object confusion. In order to patch the large hole image, we study the existing approaches and propose a new network, the compression-decompression network. The compression network takes responsibility for inpainting and generating a down-sample image. The decompression network takes responsibility for extending the down-sample image into the original resolution. We construct the compression network with the residual network and propose a similar texture selection algorithm to extend the image that is better than using the super-resolution network. We evaluate our model over Places2 and CelebA data set and use the similarity ratio as the metric. The result shows that our model has better performance when the inpainting task has many conflicts.