Medical Image Description Based on Multimodal Auxiliary Signals and Transformer

Table 2

Ablation experiments of each module.

Dataset	Methods	BLEU1	BLEU2	BLEU3	BLEU4	METEOR	ROUGE_L	CIDEr
IU-X-ray	R2Gen (base)	0.470	0.304	0.219	0.165	0.187	0.371	0.398
	MDAK	0.480	0.328	0.231	0.172	0.201	0.369	0.424
	MDAKF (add)	0.494	0.318	0.229	0.174	0.194	0.389	0.371
	MDAKF (ATT)	0.505	0.318	0.219	0.159	0.195	0.383	0.344
	MDAKF (cat)	0.484	0.307	0.221	0.167	0.192	0.391	0.334
	MDAKF (mul)	0.457	0.291	0.209	0.159	0.179	0.371	0.372
COV-CTR	R2Gen (base)	0.725	0.641	0.580	0.528	0.399	0.677	1.358
	MDAK	0.723	0.652	0.586	0.545	0.403	0.676	1.452
	MDAKF (add)	0.726	0.651	0.583	0.539	0.401	0.683	1.354
	MDAKF (ATT)	0.727	0.649	0.588	0.537	0.400	0.674	1.243
	MDAKF (cat)	0.722	0.640	0.576	0.524	0.405	0.683	1.305
	MDAKF (mul)	0.718	0.637	0.574	0.521	0.401	0.681	0.302

The bold values indicate that the model performance of the algorithm is optimal in a certain type of dataset.