Skip to main content

Table 2 Processing time (in hours) of each stage of our proposed 3D model alignment algorithm

From: Efficient video collection association using geometry-aware Bag-of-Iconics representations

Dataset Keyframe Histogram Clustering Scoring SfM Merging Total
London YouTube 227.34 11.73 6.10 132.17 4.37 2.25 383.96
Berlin YouTube 206.84 9.12 5.37 146.84 3.10 1.12 372.39
  1. SfM timing reported on top 30 video sub-sequences