Semi-supervised blockwisely architecture search for efficient lightweight generative adversarial network
作者：
Highlights：
• Using semi-supervised learning method combined with block-based architecture search, which greatly reduces the level of supervision.
• Randomly occlude a part of the picture, generate the picture according to the semantic around the occlusion block.
• The optimal architecture is constructed by flexibly stacking blocks, which realizes the image classification task with high efficiency.
• A balance is achieved between the lightweight and performance, thus the network can be well applied to mobile platform.
摘要
•Using semi-supervised learning method combined with block-based architecture search, which greatly reduces the level of supervision.•Randomly occlude a part of the picture, generate the picture according to the semantic around the occlusion block.•The optimal architecture is constructed by flexibly stacking blocks, which realizes the image classification task with high efficiency.•A balance is achieved between the lightweight and performance, thus the network can be well applied to mobile platform.
论文关键词：Semi-supervised,GANs,Network architecture search,Image generation,Image classification
论文评审过程：Received 19 February 2020, Revised 28 October 2020, Accepted 9 December 2020, Available online 17 December 2020, Version of Record 24 December 2020.
论文官网地址：https://doi.org/10.1016/j.patcog.2020.107794
