: TNO instance search submission 2012

Briefly, what approach or combination of approaches did you test in each of your submitted runs? • all runs: video decoding using ffmpeg library, sampling every 25th frame. • F_X_NO_TNO-ANNMSK_1: standard SIFT key point detection, exhaustive search using FLANN matching with query key points inside query mask. • F_X_NO_TNO-ANNMSK_3: standard SIFT key point detection, exhaustive search using FLANN matching with query key points inside query mask and resolution of query images enhanced by a factor 2 for small objects. • F_X_NO_TNO-INOMSK_2: standard SIFT key point detection within mask region, extended bag-of-words using 1000 prototypes from videos, indexing and querying using Lucene. • F_X_NO_TNO-INORDR_4: standard SIFT key point detection, extended bag-of-words using 1000 prototypes from videos, indexing and querying using Lucene.