ORB: An efficient alternative to SIFT or SURF

TLDR

Feature matching underpins many computer vision tasks, yet existing descriptors are computationally expensive. The authors propose ORB, a fast binary descriptor derived from BRIEF that is rotation invariant and robust to noise. ORB is implemented as a binary descriptor built on BRIEF and evaluated on real‑world tasks such as object detection and smartphone patch tracking. Experiments show ORB is two orders of magnitude faster than SIFT while matching its performance in many scenarios.

Abstract

Feature matching is at the base of many computer vision problems, such as object recognition or structure from motion. Current methods rely on costly descriptors for detection and matching. In this paper, we propose a very fast binary descriptor based on BRIEF, called ORB, which is rotation invariant and resistant to noise. We demonstrate through experiments how ORB is at two orders of magnitude faster than SIFT, while performing as well in many situations. The efficiency is tested on several real-world applications, including object detection and patch-tracking on a smart phone.