Train rolling stock video segmentation and classification for bogie part inspection automation: a deep learning approach

Table 5 Experimental results showing defect identification abilities of TRSE automation models with mAP as the performance indicator

Baseline methods/parameters	SSD	R-CNN	Fast R-CNN	Faster R-CNN	Yolo v1	Yolo v2	Yolo v2 with skip	Yolo v2 bifold skip	B-MHAC
mPA	0.4852	0.5325	0.5289	0.5475	0.6125	0.6589	0.6895	0.8745	0.9135
mFI	0.5987	0.5847	0.5245	0.5125	0.4528	0.4753	0.4236	0.2698	0.2258
mNI	0.5463	0.5126	0.5247	0.5169	0.4863	0.4236	0.4198	0.2891	0.2122