基于跨模态级联扩散模型的图像描述方法

陈巧红,郭孟浩,方贤,孙麒

Image captioning based on cross-modal cascaded diffusion model

Qiaohong CHEN,Menghao GUO,Xian FANG,Qi SUN

表 6 Flickr30k 数据集中不同图像描述模型的性能对比

Tab.6 Performance comparison of different image description models in Flickr30k dataset

模型 B@1 B@4 M C Deep VS[26] 57.3 15.7 15.3 24.7 Soft-Attention[2] 66.7 19.1 18.5 — Hard-Attention[2] 66.9 19.9 18.5 — Adaptive[42] 67.7 25.1 20.4 53.1 NBT[34] 69.0 27.1 21.7 57.5 Relation-Context[3] 73.6 30.1 23.8 60.2 LSTNet[43] 67.1 23.3 20.4 64.5 本研究 74.5 31.2 23.9 65.4