Chinese Journal of Mechanical Engineering

Table 4 Comparison of classification accuracies (%) on Cityscapes and KITTI datasets

From: ML-ANet: A Transfer Learning Approach Using Adaptation Network for Multi-label Image Classification in Autonomous Driving

Method	C⟶K				K⟶C
Method	Pedestrian	Vehicle	2-wheeler	Avg.	Pedestrian	Vehicle	2-wheeler	Avg.
ML-KNN [7]	62.9	94.7	54.2	70.58	56.2	74.7	56.2	62.38
TCA [16]	59.5	89.6	50.7	66.61	69.5	71.8	43.4	61.57
JDA [17]	65.7	87.1	52.7	68.49	67.3	74.5	44.2	62.03
BDA [18]	71.3	92.0	48.6	70.61	70.0	74.3	45.6	63.27
DDC [23]	77.1	93.4	55.5	75.30	73.5	80.4	52.1	68.67
DAN [31]	77.4	93.6	58.4	76.47	74.4	81.2	51.6	69.07
Ours (ML-ANet)	81.7	94.9	59.4	78.67	75.2	83.8	51.3	70.10

Back to article page