Automatic Interlace or Progressive Video Discrimination

Manish Pindoria, Tim Borer

Video content originates from a wide variety of sources. Even within one programme, several different video technologies may have been used during production. This paper discusses an algorithm that is able to reliably identify progressive and interlace frames. The algorithm is based on calculating a metric based on the degree of “interlacing artefacts” produced when adjacent fields from different frames are re-interleaved to reform a frame. The metrics are analysed over multiple frames to detect whether the material originates from a progressive or interlace source. This process has successfully been adapted to correct film-phase errors found in telecined archive material.

Published
2012-10
Content type
Original Research
Keywords
Film-mode detector, telecine
DOI
10.5594/M001440
ISBN
978-1-61482-952-2