English (United Kingdom)

https://curated-unify.zendy.io/wp-json/zendy-region/v1/featured_content/oa?rat=en

https://curated-unify.zendy.io/wp-json/zendy-region/v1/highlighted_journal/

Global: Zendy Plus annual

Presents the access of premium content as premium feature

Premium Content

Presents the keyphrase highlighting as premium feature

Keyphrase Highlighting

Presents the summarisation as premium feature

Summarisation

Insights

Presents the pdf analysis as premium feature

PDF Analysis

Presents the zaia usage as premium feature

ZAIA

Global: Zendy Plus monthly

Global: Zendy Tools annual

Global: Zendy Tools monthly

Global: Zendy Free Plan

Image inpainting aims to fill the missing hole of the input. It is hard tosolve this task efficiently when facing high-resolution images due to tworeasons: (1) Large reception field needs to be handled for high-resolutionimage inpainting. (2) The general encoder and decoder network synthesizes manybackground pixels synchronously due to the form of the image matrix. In thispaper, we try to break the above limitations for the first time thanks to therecent development of continuous implicit representation. In detail, wedown-sample and encode the degraded image to produce the spatial-adaptiveparameters for each spatial patch via an attentional Fast FourierConvolution(FFC)-based parameter generation network. Then, we take theseparameters as the weights and biases of a series of multi-layerperceptron(MLP), where the input is the encoded continuous coordinates and theoutput is the synthesized color value. Thanks to the proposed structure, weonly encode the high-resolution image in a relatively low resolution for largerreception field capturing. Then, the continuous position encoding will behelpful to synthesize the photo-realistic high-frequency textures byre-sampling the coordinate in a higher resolution. Also, our framework enablesus to query the coordinates of missing pixels only in parallel, yielding a moreefficient solution than the previous methods. Experiments show that theproposed method achieves real-time performance on the 2048$\times$2048 imagesusing a single GTX 2080 Ti GPU and can handle 4096$\times$4096 images, withmuch better performance than existing state-of-the-art methods visually andnumerically. The code is available at:https://github.com/NiFangBaAGe/CoordFill.

CoordFill: Efficient High-Resolution Image Inpainting via Parameterized Coordinate Querying