Weak label extraction from subtitle-transcript matching for person identification in TV series
Switch branches/tags
Nothing to show
Clone or download
Makarand Tapaswi
Makarand Tapaswi add a readme
Latest commit ce26a3e Mar 12, 2015
Failed to load latest commit information.
data dump all data (tracks, shots, threading clusters, subtt-trans alignme… Mar 12, 2015
initializers add initializers functions Mar 12, 2015
utilities add utils and helper functions Mar 12, 2015
.gitignore add gitignore Mar 12, 2015
README.md add a readme Mar 12, 2015
first_init.m add startup, first-init Mar 12, 2015


Weak Labels for PersonID in TV series

This is a Matlab implementation of the paper:

Improved Weak Labels using Contextual Cues for Person Identification in Videos
Makarand Tapaswi, Martin Bäuml, and Rainer Stiefelhagen
IEEE International Conference on Automatic Face and Gesture Recognition (FG), 2015
Paper download | ShotThreading & SceneDetection code

Tested on

Ubuntu 14.04 with Matlab version R2014a - R2015a.

First initialization

The first_init.m script will be called on running startup.m the first time. This will ask you to install some external toolboxes. Please follow the instructions.

Example usage

A video structure can be created by calling either
VS = BBT(1, 1); or VS = BUFFY(5, 1)

The main function can be directly invoked with one or multiple videos at once. ft = speaking_face2_wrapper(BBT(1, 1:6));

We include data for 6 episodes each of

  • The Big Bang Theory (BBT) season 1, episodes 1..6
  • Buffy the Vampire Slayer (BUFFY) season 5, episodes 1..6 as described in the paper.

For any questions about the generation of data, please contact me.

External toolboxes

Main functions


  • 12-03-2015: A complete working example of the FG 2015 paper