TY - GEN
T1 - Bottom-up and top-down attention for image captioning and visual question answering
AU - Anderson, Peter
AU - He, Xiaodong
AU - Buehler, Chris
AU - Teney, Damien
AU - Johnson, Mark
AU - Gould, Stephen
AU - Zhang, Lei
PY - 2018/12/14
Y1 - 2018/12/14
N2 - Top-down visual attention mechanisms have been used extensively in image captioning and visual question answering (VQA) to enable deeper image understanding through fine-grained analysis and even multiple steps of reasoning. In this work, we propose a combined bottom-up and top-down attention mechanism that enables attention to be calculated at the level of objects and other salient image regions. This is the natural basis for attention to be considered. Within our approach, the bottom-up mechanism (based on Faster R-CNN) proposes image regions, each with an associated feature vector, while the top-down mechanism determines feature weightings. Applying this approach to image captioning, our results on the MSCOCO test server establish a new state-of-the-art for the task, achieving CIDEr / SPICE / BLEU-4 scores of 117.9, 21.5 and 36.9, respectively. Demonstrating the broad applicability of the method, applying the same approach to VQA we obtain first place in the 2017 VQA Challenge.
AB - Top-down visual attention mechanisms have been used extensively in image captioning and visual question answering (VQA) to enable deeper image understanding through fine-grained analysis and even multiple steps of reasoning. In this work, we propose a combined bottom-up and top-down attention mechanism that enables attention to be calculated at the level of objects and other salient image regions. This is the natural basis for attention to be considered. Within our approach, the bottom-up mechanism (based on Faster R-CNN) proposes image regions, each with an associated feature vector, while the top-down mechanism determines feature weightings. Applying this approach to image captioning, our results on the MSCOCO test server establish a new state-of-the-art for the task, achieving CIDEr / SPICE / BLEU-4 scores of 117.9, 21.5 and 36.9, respectively. Demonstrating the broad applicability of the method, applying the same approach to VQA we obtain first place in the 2017 VQA Challenge.
UR - http://www.scopus.com/inward/record.url?scp=85053519533&partnerID=8YFLogxK
U2 - 10.1109/CVPR.2018.00636
DO - 10.1109/CVPR.2018.00636
M3 - Conference proceeding contribution
AN - SCOPUS:85053519533
SN - 9781538664216
T3 - IEEE Conference on Computer Vision and Pattern Recognition
SP - 6077
EP - 6086
BT - Proceedings - 2018 IEEE/CVF Conference on Computer Vision and Pattern Recognition, CVPR 2018
PB - Institute of Electrical and Electronics Engineers (IEEE)
CY - Piscataway, NJ
T2 - 31st Meeting of the IEEE/CVF Conference on Computer Vision and Pattern Recognition, CVPR 2018
Y2 - 18 June 2018 through 22 June 2018
ER -