Aggregation of binary feature descriptors for compact scene model\n representation in large scale structure-from-motion applications

2018/09/28 by Jacek Komorowski, Komorowski, Jacek, T. P. Trzcinski +1
Computer Science · Engineering · #Advanced Image and Video Retrieval Techniques #Advanced Vision and Imaging #Computer Vision and Pattern Recognition (cs.CV) #FOS: Computer and information sciences #Robotics and Sensor-Based Localization

paper · pdf · doi:10.48550/arxiv.1809.11062

openalex publication_date 2018/09/28 · openalex created_date 2022/08/02 · openalex updated_date 2026/07/28

Abstract

In this paper we present an efficient method for aggregating binary feature\ndescriptors to allow compact representation of 3D scene model in incremental\nstructure-from-motion and SLAM applications. All feature descriptors linked\nwith one 3D scene point or landmark are represented by a single low-dimensional\nreal-valued vector called a \prototype. The method allows significant\nreduction of memory required to store and process feature descriptors in\nlarge-scale structure-from-motion applications. An efficient approximate\nnearest neighbours search methods suited for real-valued descriptors, such as\nFLANN, can be used on the resulting prototypes to speed up matching processed\nframes.\n

Citations

• Adam: A Method for Stochastic Optimization
• Distinctive Image Features from Scale-Invariant Keypoints

Related

• ProLFA: Representative Prototype Selection for Local Feature Aggregation
• Hyperdimensional computing as a framework for systematic aggregation of\n image descriptors
• FeatureBooster: Boosting Feature Descriptors with a Lightweight Neural Network
• CodeSLAM - Learning a Compact, Optimisable Representation for Dense Visual SLAM
• Integrating Objects into Monocular SLAM: Line Based Category Specific Models