Object detection in high resolution images based on multiscale and block processing
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 lay...
Main Authors: | , , |
---|---|
Format: | Article |
Language: | Russian |
Published: |
The United Institute of Informatics Problems of the National Academy of Sciences of Belarus
2020-06-01
|
Series: | Informatika |
Subjects: | |
Online Access: | https://inf.grid.by/jour/article/view/1056 |
_version_ | 1797877214160617472 |
---|---|
author | R. P. Bohush I. Yu. Zakharava S. V. Ablameyko |
author_facet | R. P. Bohush I. Yu. Zakharava S. V. Ablameyko |
author_sort | R. P. Bohush |
collection | DOAJ |
description | 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. |
first_indexed | 2024-04-10T02:13:38Z |
format | Article |
id | doaj.art-15037df4de9248b2b7f6161ade40a3ae |
institution | Directory Open Access Journal |
issn | 1816-0301 |
language | Russian |
last_indexed | 2024-04-10T02:13:38Z |
publishDate | 2020-06-01 |
publisher | The United Institute of Informatics Problems of the National Academy of Sciences of Belarus |
record_format | Article |
series | Informatika |
spelling | doaj.art-15037df4de9248b2b7f6161ade40a3ae2023-03-13T08:32:24ZrusThe United Institute of Informatics Problems of the National Academy of Sciences of BelarusInformatika1816-03012020-06-0117271610.37661/1816-0301-2020-17-2-7-16925Object detection in high resolution images based on multiscale and block processingR. P. Bohush0I. Yu. Zakharava1S. V. Ablameyko2Polotsk State UniversityPolotsk State UniversityBelarusian State University; The United Institute of Informatics Problems of the National Academy of Sciences of BelarusIn 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.https://inf.grid.by/jour/article/view/1056convolutional neural networkblock processing4к resolutionobject detectionmultiscale image representation |
spellingShingle | R. P. Bohush I. Yu. Zakharava S. V. Ablameyko Object detection in high resolution images based on multiscale and block processing Informatika convolutional neural network block processing 4к resolution object detection multiscale image representation |
title | Object detection in high resolution images based on multiscale and block processing |
title_full | Object detection in high resolution images based on multiscale and block processing |
title_fullStr | Object detection in high resolution images based on multiscale and block processing |
title_full_unstemmed | Object detection in high resolution images based on multiscale and block processing |
title_short | Object detection in high resolution images based on multiscale and block processing |
title_sort | object detection in high resolution images based on multiscale and block processing |
topic | convolutional neural network block processing 4к resolution object detection multiscale image representation |
url | https://inf.grid.by/jour/article/view/1056 |
work_keys_str_mv | AT rpbohush objectdetectioninhighresolutionimagesbasedonmultiscaleandblockprocessing AT iyuzakharava objectdetectioninhighresolutionimagesbasedonmultiscaleandblockprocessing AT svablameyko objectdetectioninhighresolutionimagesbasedonmultiscaleandblockprocessing |