Connected Components-based Colour Image Representations of Vibrations for a Two-stage Fault Diagnosis of Roller Bearings Using Convolutional Neural Networks

Chinese Journal of Mechanical Engineering

Table 7 A comparison with the classification results from the literature on bearing datasets A, B, and C datasets

Reference No.	Method	Dataset	Testing data size (%)	Classification accuracy (%)
[34]	DNN	A	50	99.95 ± 0.06
		B		99.61 ± 0.21
		C		99.74 ± 0.16
	BPNN	A	50	65.2 ± 18.09
		B		61.95 ± 22.09
		C		69.82 ± 17.67
[36]	CS-DNN with CS sampling rate of 0.1	A	50	99.3 ± 0.6
		B		99.7 ± 0.5
		C		100 ± 0.0
[91]	MLP	A	50	95.7
		B		99.6
		C		99.4
[67]	AlexNet	\(A\to B;A\to C;B\to A;\)	–	94.3
	ResNet	\(B\to C;C\to A;C\to B\)		94.6
	ICN	\(B\to C;C\to A;C\to B\)		97.2
This paper	RGBVI-CNN *the proposed method	A	50	99.6 ± 0.4
		B		99.2 ± 0.7
		C		99.3± 0.7
		A	40	99.9 ± 0.1
		B		99.8 ± 0.1
		C		99.6± 0.3
		A	30	100 ± 0.0
		B		99.9 ± 0.1
		C		99.9± 0.1