Open AccessUSING MORPHLET-BASED IMAGE REPRESENTATION FOR OBJECT DETECTION
Author(s) -
V. S. Gorbatsevich,
Yu. V. Vizilter
Publication year - 2016
Publication title -
the international archives of the photogrammetry, remote sensing and spatial information sciences/international archives of the photogrammetry, remote sensing and spatial information sciences
Language(s) - English
Resource type - Journals
SCImago Journal Rank - 0.264
H-Index - 71
eISSN - 1682-1777
pISSN - 1682-1750
DOI - 10.5194/isprsarchives-xli-b3-859-2016
Subject(s) - artificial intelligence , computer science , pyramid (geometry) , bounding overwatch , representation (politics) , computer vision , object detection , image (mathematics) , pattern recognition (psychology) , feature detection (computer vision) , feature (linguistics) , segmentation , minimum bounding box , tree (set theory) , window (computing) , image segmentation , object (grammar) , image processing , mathematics , mathematical analysis , linguistics , philosophy , geometry , politics , political science , law , operating system
In this paper, we propose an original method for objects detection based on a special tree-structured image representation – the trees of morphlets. The method provides robust detection of various types of objects in an image without employing a machine learning procedure. Along with a bounding box creation on a detection step, the method makes pre-segmentation, which can be further used for recognition purposes. Another important feature of the proposed approach is that there are no needs to use a running window as well as a features pyramid in order to detect the objects of different sizes.