Bags of Binary Words for Fast Recognition in Image Sequences
lastmod 2026-11-10created 2019-09-265 edits
Bags of Binary Words for Fast Recognition in Image Sequences is a paper by Gálvez-López and Tardós that proposes grouping binary descriptors such as BRIEF and ORB into a vocabulary to quickly determine the similarity between images. Because it compares images converted into word-frequency vectors, it can perform place recognition (loop closure detection) in real time. DBoW2, the implementation of this method, is used as-is for loop detection in the ORB-SLAM family.
Documentation
- Bags of Binary Words for Fast Recognition in Image Sequences
- http://webdiis.unizar.es/~dorian/papers/GalvezTRO12.pdf
-
GalvezTRO12.pdf