Title :
Memory Efficient 3D Integral Volumes
Author :
Urschler, Martin ; Bornik, Alexander ; Donoser, Michael
Author_Institution :
Ludwig Boltzmann Inst. for Clinical Forensic Imaging, Graz, Austria
Abstract :
Integral image data structures are very useful in computer vision applications that involve machine learning approaches based on ensembles of weak learners. The weak learners often are simply several regional sums of intensities subtracted from each other. In this work we present a memory efficient integral volume data structure, that allows reduction of required RAM storage size in such a supervised learning framework using 3D training data. We evaluate our proposed data structure in terms of the tradeoff between computational effort and storage, and show an application for 3D object detection of liver CT data.
Keywords :
computer vision; computerised tomography; data structures; learning (artificial intelligence); liver; medical image processing; object detection; random-access storage; 3D object detection; 3D training data; RAM storage size reduction; computer vision; integral image data structures; integral volume data structure; liver CT data; machine learning; memory efficient 3D integral volumes; supervised learning framework; Computational modeling; Computed tomography; Data structures; Memory management; Solid modeling; Three-dimensional displays; Training; integral volume; memory efficient; object detection; random forest; summed volume table;
Conference_Titel :
Computer Vision Workshops (ICCVW), 2013 IEEE International Conference on
Conference_Location :
Sydney, NSW
DOI :
10.1109/ICCVW.2013.99