id	author	title	date	pages	extension	mime	words	sentence	flesch	summary	cache	txt
fcis-1973	He, Jingyuan; Yang, Bailong	Fine-grained Image Recognition Method using Discriminative Region-based Data Augmentation	2022	5	.pdf	application/pdf	4045	195	46	The random cropping data augmentation method proposed by Krizhevsky et al. can effectively improve the accuracy, and then a series of image translation data augmentation methods appear. Drawing on the idea of data augmentation algorithm based on attention mechanism, literature proposes a multi-sample data augmentation algorithm based on spatial and channel attention for fine-grained image classification, which can effectively distinguish image features while making the training data more diverse, and further solve the problem of overfitting of network models.	cache/fcis-1973.pdf	txt/fcis-1973.txt