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 Tools annual

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 monthly

Global: Zendy Plus monthly

Presents the access of premium content as premium feature

Premium Content

Global: Zendy Plus annual

Global: Zendy Free Plan

The effective extraction of spatial-angular features plays a crucial role inlight field image super-resolution (LFSR) tasks, and the introduction ofconvolution and Transformers leads to significant improvement in this area.Nevertheless, due to the large 4D data volume of light field images, manyexisting methods opted to decompose the data into a number of lower-dimensionalsubspaces and perform Transformers in each sub-space individually. As a sideeffect, these methods inadvertently restrict the self-attention mechanisms to aOne-to-One scheme accessing only a limited subset of LF data, explicitlypreventing comprehensive optimization on all spatial and angular cues. In thispaper, we identify this limitation as subspace isolation and introduce a novelMany-to-Many Transformer (M2MT) to address it. M2MT aggregates angularinformation in the spatial subspace before performing the self-attentionmechanism. It enables complete access to all information across allsub-aperture images (SAIs) in a light field image. Consequently, M2MT isenabled to comprehensively capture long-range correlation dependencies. WithM2MT as the pivotal component, we develop a simple yet effective M2MT networkfor LFSR. Our experimental results demonstrate that M2MT achievesstate-of-the-art performance across various public datasets. We further conductin-depth analysis using local attribution maps (LAM) to obtain visualinterpretability, and the results validate that M2MT is empowered with a trulynon-local context in both spatial and angular subspaces to mitigate subspaceisolation and acquire effective spatial-angular representation.

Beyond Subspace Isolation: Many-to-Many Transformer for Light Field Image Super-resolution