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 Tools annual

Global: Zendy Free Plan

Global: Zendy Tools monthly

Global: Zendy Plus monthly

This paper presents an efficient deep-learning model named Swin Transformer fusion network (STFN) for full-reference image quality assessment (FR-IQA). The STFN model uses the first and second stages of the Swin Transformer for feature extraction. To unify the features from these two stages, we propose fusion operations including reverse patch merging (RPM) and mediator block (MB) operations. The RPM is a kind of reverse operation of the patch merging operation in the Swin Transformer stage, and it reshapes the size of the second stage feature so as to match that of the first stage feature. The MB operation efficiently combines multiple features from the RPM block and the first stage Swin Transformer for subsequent operations. Experimental results show that the proposed STFN model provides significantly improved performance than the previous traditional and deep-learning models for various kinds of image datasets for FR-IQA. The STFN model also shows superior performance compared to the state-of-the-art method for FR-IQA with smaller training time and model size. The code and pretrained models are publicly available at https://github.com/KIIPLab/STFN.

Swin Transformer Fusion Network for Image Quality Assessment