id	author	title	date	pages	extension	mime	words	sentence	flesch	summary	cache	txt
ajst-16785	Tang, Donglin; Zhao, Yunliang	Inception meets Swin Transformer: A Novel Approach for Metal Defect Recognition	2024	10	.pdf	application/pdf	6358	328	49	In order to verify the effectiveness of ICST for the recognition of metal defect ultrasonic signal grayscale images, a self-built ultrasonic detection defect dataset (ULFSL-DET) was constructed. In contrast, ICST achieves the detection of metal defect ultrasonic signal grayscale images and image-based mental defects with lower computation cost, while obtaining optimal accuracy in small and medium-sized datasets.	cache/ajst-16785.pdf	txt/ajst-16785.txt