MOCHA: Multi-modal Objects-aware Cross-arcHitecture Alignment

Published in: Proceedings of the European Conference on Computer Vision (ECCV)