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/

Zendy Plus

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

Zendy Tools

Zendy Open

Deep neural networks for image super-resolution (SR) have demonstratedsuperior performance. However, the large memory and computation consumptionhinders their deployment on resource-constrained devices. Binary neuralnetworks (BNNs), which quantize the floating point weights and activations to1-bit can significantly reduce the cost. Although BNNs for image classificationhave made great progress these days, existing BNNs for SR still suffer from alarge performance gap between the FP SR networks. To this end, we observe theactivation distribution in SR networks and find much larger pixel-to-pixel,channel-to-channel, layer-to-layer, and image-to-image variation in theactivation distribution than image classification networks. However, existingBNNs for SR fail to capture these variations that contain rich information forimage reconstruction, leading to inferior performance. To address this problem,we propose SCALES, a binarization method for SR networks that consists of thelayer-wise scaling factor, the spatial re-scaling method, and the channel-wisere-scaling method, capturing the layer-wise, pixel-wise, and channel-wisevariations efficiently in an input-dependent manner. We evaluate our methodacross different network architectures and datasets. For CNN-based SR networks,our binarization method SCALES outperforms the prior art method by 0.2dB withfewer parameters and operations. With SCALES, we achieve the first accuratebinary Transformer-based SR network, improving PSNR by more than 1dB comparedto the baseline method.

SCALES: Boost Binary Neural Network for Image Super-Resolution with Efficient Scalings