Abstract | ||
---|---|---|
This paper demonstrates a novel method for automatically discovering and recognising characters in video without any labelled examples or user intervention. Instead weak supervision is obtained via a rough script-to-subtitle alignment. The technique uses pose invariant features, extracted from detected faces and clustered to form groups of co-occurring characters. Results show that with 9 characters, 29% of the closest exemplars are correctly identified, increasing to 50% as additional exemplars are considered. |
Year | DOI | Venue |
---|---|---|
2014 | 10.1007/978-3-319-13737-7_11 | Lecture Notes in Computer Science |
Field | DocType | Volume |
Computer vision,Linear prediction,Active appearance model,Artificial intelligence,Invariant (mathematics),Mathematics | Conference | 8912 |
ISSN | Citations | PageRank |
0302-9743 | 1 | 0.35 |
References | Authors | |
27 | 3 |
Name | Order | Citations | PageRank |
---|---|---|---|
Matthew Marter | 1 | 1 | 0.35 |
Simon Hadfield | 2 | 356 | 22.65 |
Richard Bowden | 3 | 1840 | 118.50 |