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

Physical photographs now can be conveniently scanned by smartphones andstored forever as a digital version, yet the scanned photos are not restoredwell. One solution is to train a supervised deep neural network on many digitalphotos and the corresponding scanned photos. However, it requires a high laborcost, leading to limited training data. Previous works create training pairs bysimulating degradation using image processing techniques. Their syntheticimages are formed with perfectly scanned photos in latent space. Even so, thereal-world degradation in smartphone photo scanning remains unsolved since itis more complicated due to lens defocus, lighting conditions, losing detailsvia printing. Besides, locally structural misalignment still occurs in data dueto distorted shapes captured in a 3-D world, reducing restoration performanceand the reliability of the quantitative evaluation. To solve these problems, wepropose a semi-supervised Deep Photo Scan (DPScan). First, we present a way ofproducing real-world degradation and provide the DIV2K-SCAN dataset forsmartphone-scanned photo restoration. Also, Local Alignment is proposed toreduce the minor misalignment remaining in data. Second, we simulate manydifferent variants of the real-world degradation using low-level imagetransformation to gain a generalization in smartphone-scanned image properties,then train a degradation network to generalize all styles of degradation andprovide pseudo-scanned photos for unscanned images as if they were scanned by asmartphone. Finally, we propose a Semi-Supervised Learning that allows ourrestoration network to be trained on both scanned and unscanned images,diversifying training image content. As a result, the proposed DPScanquantitatively and qualitatively outperforms its baseline architecture,state-of-the-art academic research, and industrial products in smartphone photoscanning.

Deep Photo Scan: Semi-Supervised Learning for dealing with the real-world degradation in Smartphone Photo Scanning