• shareshare
  • link
  • cite
  • add
auto_awesome_motion View all 2 versions
Publication . Article . 2020

Object detection in high resolution images based on multiscale and block processing

Rykhard Bohush; I. Yu. Zakharava; Sergey Ablameyko;
Open Access
Published: 26 Jun 2020 Journal: Informatics, volume 17, pages 7-16 (issn: 1816-0301, eissn: 2617-6963, Copyright policy )
Publisher: United Institute of Informatics Problems of the National Academy of Sciences of Belarus
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.
Subjects by Vocabulary

Microsoft Academic Graph classification: Block (data storage) Object detection Computer science Image representation Pattern recognition Image resolution Layer (object-oriented design) Convolutional neural network Artificial intelligence business.industry business Pyramid (image processing) High resolution

ACM Computing Classification System: ComputingMethodologies_IMAGEPROCESSINGANDCOMPUTERVISION


General Medicine, convolutional neural network, block processing, 4к resolution, object detection, multiscale image representation, Electronic computers. Computer science, QA75.5-76.95

Related Organizations