ViP-CNN: Visual Phrase Guided Convolutional Neural Network

Abstract

Our most recent works on Visual Relationship Detection will be presented during CVPR2017 at poster session. We propose an end-to-end trained relationship recognition network, which outperforms the state-of-the-art method by a large margin.

Date
Location
Kamehameha I, Hawaii Convention Center, Honolulu, Hawaii, U.S.
Links