arxivst stuff from arxiv that you should probably bookmark

Detecting and Recognizing Human-Object Interactions

Abstract · Apr 24, 2017 17:14 ·

centric interactions object person detected human triplets action cs-cv

Arxiv Abstract

  • Georgia Gkioxari
  • Ross Girshick
  • Piotr Dollár
  • Kaiming He

To understand the visual world, a machine must not only recognize individual object instances but also how they interact. Humans are often at the center of such interactions and detecting human-object interactions is an important practical and scientific problem. In this paper, we address the task of detecting triplets in challenging everyday photos. We propose a novel model that is driven by a human-centric approach. Our hypothesis is that the appearance of a person – their pose, clothing, action – is a powerful cue for localizing the objects they are interacting with. To exploit this cue, our model learns to predict an action-specific density over target object locations based on the appearance of a detected person. Our model also jointly learns to detect people and objects, and by fusing these predictions it efficiently infers interaction triplets in a clean, jointly trained end-to-end system we call Interact. We validate our approach on the recently introduced Verbs in COCO (V-COCO) dataset, where we show qualitatively and quantitatively compelling results.

Read the paper (pdf) »