Semi-supervised blockwisely architecture search for efficient lightweight generative adversarial network. (April 2021)

• Record Type: Journal Article
• Authors: Zhang, Man; Zhou, Yong; Zhao, Jiaqi; Xia, Shixiong; Wang, Jiaqi; Huang, Zizheng
• Is Part Of: Pattern recognition. Volume 112(2021)
• Publication Date: 2021-04
• Subjects: Semi-supervised -- GANs -- Network architecture search -- Image generation -- Image classification; 00-01 -- 99-00; Pattern perception -- Periodicals; Perception des structures -- Périodiques; Patroonherkenning; 006.4
• DOI: 10.1016/j.patcog.2020.107794
• Languages: English
• ISSNs: 0031-3203

Abstract

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. Abstract: In the field of computer vision, methods that use fully supervised learning and fixed deep network structures need to be improved. Currently, many studies are devoted to designing neural architecture search methods to use neural networks in a more flexible way. However, most of these methods use fully supervised learning at the cost of extraordinary GPU training time. In view of the above problems, we propose a semi-supervised generative adversarial network and search network architecture based on block structure. Use real pictures and generated pictures with corresponding real tags and pseudo tags for training, to achieve the purpose of semi-supervised learning. By setting the layer's hyperparameters to a variable and flexible stacking block structure, network architecture search is achieved. The proposed method realizes image generation and extends to image classification. In the experimental results in Section 4, the training time is greatly reduced and the model performance is improved, which illustrates the efficiency of our method. The code can be found in https://github.com/AICV-CUMT/STASGAN .