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

Recent work showed neural-network-based approaches to reconstructing imagesfrom compressively sensed measurements offer significant improvements inaccuracy and signal compression. Such methods can dramatically boost thecapability of computational imaging hardware. However, to date, there have beentwo major drawbacks: (1) the high-precision real-valued sensing patternsproposed in the majority of existing works can prove problematic when used withcomputational imaging hardware such as a digital micromirror sampling deviceand (2) the network structures for image reconstruction involve intensivecomputation, which is also not suitable for hardware deployment. To addressthese problems, we propose a novel hardware-friendly solution based onmixed-weights neural networks for computational imaging. In particular, learnedbinary-weight sensing patterns are tailored to the sampling device. Moreover,we proposed a recursive network structure for low-resolution image sampling andhigh-resolution reconstruction scheme. It reduces both the required number ofmeasurements and reconstruction computation by operating convolution on smallintermediate feature maps. The recursive structure further reduced the modelsize, making the network more computationally efficient when deployed with thehardware. Our method has been validated on benchmark datasets and achieved thestate of the art reconstruction accuracy. We tested our proposed network inconjunction with a proof-of-concept hardware setup.

LSHR-Net: a hardware-friendly solution for high-resolution computational imaging using a mixed-weights neural network