Publications by authors named "Shengzhong Mao"
Entropy (Basel)
September 2022
Article Synopsis
- The text discusses advancements in computer vision, highlighting the popularity of deep convolutional neural networks (CNNs) and the challenge of increased resource demand when stacking layers.
- It addresses the limitations of resource-intensive models in scenarios with constrained hardware and proposes a new architecture called Universal Pixel Attention Networks (UPANets) that enhances performance efficiently.
- UPANets incorporate channel and spatial direction attention mechanisms, allowing them to learn global information with lower resource consumption while outperforming many state-of-the-art models on CIFAR datasets.
