Mid-level Features Improve Recognition of Interactive Activities

Abstract

We argue that mid-level representations can bridge the gap between existing low-level models, which are incapable of capturing the structure of interactive verbs, and contemporary high-level schemes, which rely on the output of potentially brittle intermediate detectors and trackers. We develop a novel descriptor based on generic object foreground segments our representation forms a histogram-of-gradient representation that is grounded to the frame of detected key-segments. Importantly, our method does not require objects to be identi ed reliably in order to compute a ro- bust representation. We evaluate an integrated system including novel key-segment activity descriptors on a large-scale video dataset containing 48 common verbs, for which we present a comprehensive evaluation protocol. Our results con rm that a descriptor de ned on mid-level primitives operating at a higher-level than local spatio-temporal features, but at a lower-level than trajectories of detected objects, can provide a substantial improvement relative to either alone or to their combination.

Open PDF

Document Details

Document Type
Technical Report
Publication Date
Nov 14, 2012
Accession Number
ADA570728

Entities

People

  • Ben Packer
  • Chun-Hui Chen
  • Daniel Koller
  • Fei-Fei Li
  • J. Niebles
  • K. Grauman
  • Kate Saenko
  • S. Bandla
  • Trevor Darrell
  • Y. Lee
  • Yangqing Jia

Organizations

  • University of California, Berkeley

Tags

DTIC Thesaurus Topics

  • Agreements
  • Artificial Intelligence
  • Bernoulli Distribution
  • Computer Science
  • Computer Vision
  • Detection
  • Detectors
  • Displacement
  • Electrical Engineering
  • Histograms
  • Object Recognition
  • Probability
  • Random Variables
  • Recognition
  • Test And Evaluation
  • Trajectories
  • Video Clips

Fields of Study

  • Computer science

Readers

  • Computational Linguistics
  • Distributed Systems and Data Platform Development
  • Systems Analysis and Design