Open AccessObject detection in high resolution images based on multiscale and block processing
Author(s) -
Rykhard Bohush,
I. Yu. Zakharava,
Sergey Ablameyko
Publication year - 2020
Publication title -
informatika
Language(s) - English
Resource type - Journals
eISSN - 2617-6963
pISSN - 1816-0301
DOI - 10.37661/1816-0301-2020-17-2-7-16
Subject(s) - pyramid (geometry) , block (permutation group theory) , artificial intelligence , computer science , object detection , convolutional neural network , pattern recognition (psychology) , representation (politics) , object (grammar) , image (mathematics) , image processing , layer (electronics) , computer vision , artificial neural network , mathematics , chemistry , geometry , organic chemistry , politics , political science , law
In the paper the algorithm for object detection in high resolution images is proposed. The approach uses multiscale image representation followed by block processing with the overlapping value. For each block the object detection with convolutional neural network was performed. Number of pyramid layers is limited by the Convolutional Neural Network layer size and input image resolution. Overlapping blocks splitting to improve the classification and detection accuracy is performed on each layer of pyramid except the highest one. Detected areas are merged into one if they have high overlapping value and the same class. Experimental results for the algorithm are presented in the paper.