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

Convolutional neural networks (CNNs) have enabled significant performanceleaps in medical image classification tasks. However, translating neuralnetwork models for clinical applications remains challenging due to dataprivacy issues. Fully Homomorphic Encryption (FHE) has the potential to addressthis challenge as it enables the use of CNNs on encrypted images. However,current HE technology poses immense computational and memory overheads,particularly for high-resolution images such as those seen in the clinicalcontext. We present CaRENets: Compact and Resource-Efficient CNNs for highperformance and resource-efficient inference on high-resolution encryptedimages in practical applications. At the core, CaRENets comprises a new FHEcompact packing scheme that is tightly integrated with CNN functions. CaRENetsoffers dual advantages of memory efficiency (due to compact packing of imagesand CNN activations) and inference speed (due to the reduction in the number ofciphertexts created and the associated mathematical operations) over standardinterleaved packing schemes. We apply CaRENets to perform homomorphicabnormality detection with 80-bit security level in two clinical conditions -Retinopathy of Prematurity (ROP) and Diabetic Retinopathy (DR). The ROP datasetcomprises 96 x 96 grayscale images, while the DR dataset comprises 256 x 256RGB images. We demonstrate over 45x improvement in memory efficiency and 4-5xspeedup in inference over the interleaved packing schemes. As our approachenables memory-efficient low-latency HE inference without imposing additionalcommunication burden, it has implications for practical and secure deeplearning inference in clinical imaging.

CaRENets: Compact and Resource-Efficient CNN for Homomorphic Inference on Encrypted Medical Images