Visual-Text Reference Pretraining Model for Image Captioning

Table 4

Comparison of the results generated by the visual reference network of VTR-PTM in single-channel and dual-channel coding on MSCOCO.

Approach	B@1	B@2	B@3	B@4	M	R	C	S
VRN-SC	81.9	67.1	53.2	40.7	30.3	61.0	129.7	28.2
VRN-DC	82.9	67.3	53.4	40.9	30.9	61.5	130.2	28.5