Knowledge Graph Representation Fusion Framework for Fine-Grained Object Recognition in Smart Cities

Table 2

Experimental results on Caltech-UCSD Bird dataset.

Method	Bounding box annotations	Part annotations	Accuracy ()
Part-RCNN [18]	✓	✓	76.4
SPDA-CNN [36]	✓	✓	85.14
PN-CNN [28]	✓	✓	85.4
CB-CNN [34]	✓		84.6
FCAN [21]	✓		84.7
Bilinear CNN [33]			84.1
ST-CNN [43]			84.1
CVL [46]			85.55
A3M [53]			86.2
PC-DenseNet-161 [54]			86.87
KERL [23]			87.0
SEF [55]			87.3
Ensemble T-CNN [24]			87.3
TASN [56]			87.9
HSE [57]			88.1
GCL [58]			88.3
Ours			88.49