TY - GEN
T1 - Exploring Long Tail Visual Relationship Recognition with Large Vocabulary
AU - Abdelkarim, Sherif
AU - Agarwal, Aniket
AU - Achlioptas, Panos
AU - Chen, Jun
AU - Huang, Jiaji
AU - Li, Boyang
AU - Church, Kenneth
AU - Elhoseiny, Mohamed
N1 - Publisher Copyright:
© 2021 IEEE
PY - 2021
Y1 - 2021
N2 - Several approaches have been proposed in recent literature to alleviate the long-tail problem, mainly in object classification tasks. In this paper, we make the first large-scale study concerning the task of Long-Tail Visual Relationship Recognition (LTVRR). LTVRR aims at improving the learning of structured visual relationships that come from the long-tail (e.g., “rabbit grazing on grass”). In this setup, the subject, relation, and object classes each follow a long-tail distribution. To begin our study and make a future benchmark for the community, we introduce two LTVRR-related benchmarks, dubbed VG8K-LT and GQA-LT, built upon the widely used Visual Genome and GQA datasets. We use these benchmarks to study the performance of several state-of-the-art long-tail models on the LTVRR setup. Lastly, we propose a visiolinguistic hubless (VilHub) loss and a Mixup augmentation technique adapted to LTVRR setup, dubbed as RelMix. Both VilHub and RelMix can be easily integrated on top of existing models and despite being simple, our results show that they can remarkably improve the performance, especially on tail classes. Benchmarks, code, and models have been made available at: https://github.com/Vision-CAIR/LTVRR.
AB - Several approaches have been proposed in recent literature to alleviate the long-tail problem, mainly in object classification tasks. In this paper, we make the first large-scale study concerning the task of Long-Tail Visual Relationship Recognition (LTVRR). LTVRR aims at improving the learning of structured visual relationships that come from the long-tail (e.g., “rabbit grazing on grass”). In this setup, the subject, relation, and object classes each follow a long-tail distribution. To begin our study and make a future benchmark for the community, we introduce two LTVRR-related benchmarks, dubbed VG8K-LT and GQA-LT, built upon the widely used Visual Genome and GQA datasets. We use these benchmarks to study the performance of several state-of-the-art long-tail models on the LTVRR setup. Lastly, we propose a visiolinguistic hubless (VilHub) loss and a Mixup augmentation technique adapted to LTVRR setup, dubbed as RelMix. Both VilHub and RelMix can be easily integrated on top of existing models and despite being simple, our results show that they can remarkably improve the performance, especially on tail classes. Benchmarks, code, and models have been made available at: https://github.com/Vision-CAIR/LTVRR.
UR - http://www.scopus.com/inward/record.url?scp=85122119408&partnerID=8YFLogxK
U2 - 10.1109/ICCV48922.2021.01562
DO - 10.1109/ICCV48922.2021.01562
M3 - Conference contribution
AN - SCOPUS:85122119408
T3 - Proceedings of the IEEE International Conference on Computer Vision
SP - 15901
EP - 15910
BT - Proceedings - 2021 IEEE/CVF International Conference on Computer Vision, ICCV 2021
PB - Institute of Electrical and Electronics Engineers Inc.
T2 - 18th IEEE/CVF International Conference on Computer Vision, ICCV 2021
Y2 - 11 October 2021 through 17 October 2021
ER -