arxivst stuff from arxiv that you should probably bookmark

Temporal Action Localization by Structured Maximal Sums

Abstract · Apr 15, 2017 18:10 ·

maximal evolution end temporal structured actions start frame action sum cs-cv

Arxiv Abstract

  • Zehuan Yuan
  • Jonathan C. Stroud
  • Tong Lu
  • Jia Deng

We address the problem of temporal action localization in videos. We pose action localization as a structured prediction over arbitrary-length temporal windows, where each window is scored as the sum of frame-wise classification scores. Additionally, our model classifies the start, middle, and end of each action as separate components, allowing our system to explicitly model each action’s temporal evolution and take advantage of informative temporal dependencies present in this structure. In this framework, we localize actions by searching for the structured maximal sum, a problem for which we develop a novel, provably-efficient algorithmic solution. The frame-wise classification scores are computed using features from a deep Convolutional Neural Network (CNN), which are trained end-to-end to directly optimize for a novel structured objective. We evaluate our system on the THUMOS 14 action detection benchmark and achieve competitive performance.

Read the paper (pdf) »