ILCS: An Improved Lightweight Convolution Structure and Mixed Interactive Attention for Steel Surface Defect Classification

Algorithm 1

The classification model based on ILCS.

(i)	Input: Defect image dataset , where , is the total training dataset, and represents the number of defect categories.
	Output: Defect category , where
(1)	Preprocessing of : horizontal flip, random clipping, Gaussian blur;
(2)
(3)	for in do
(4)	;
(5)	;// in equation (6)
(6)	;// in equation (4)
(7)	;// in equation (8)
(8)	= ;
(9)	= argmax ;//argmax() is a function of python
(10)	;
(11)	return Y;