Application of multi-step image segmentation for near-duplicate image recognition
The urgency of the paper is caused by the need to detect image near-duplicate in computer vision systems, as well as when image searching on Internet or in large digital archives. The main aim of the study: usage of multi-step segmentation for near-duplicate image recognition. The methods used in th...
Main Authors: | , |
---|---|
Format: | Article |
Language: | Russian |
Published: |
Tomsk Polytechnic University
2019-05-01
|
Series: | Известия Томского политехнического университета: Инжиниринг георесурсов |
Subjects: | |
Online Access: | http://izvestiya.tpu.ru/archive/article/view/1371 |
_version_ | 1797813238528737280 |
---|---|
author | Victor Nemirovskiy Alexander Stoyanov |
author_facet | Victor Nemirovskiy Alexander Stoyanov |
author_sort | Victor Nemirovskiy |
collection | DOAJ |
description | The urgency of the paper is caused by the need to detect image near-duplicate in computer vision systems, as well as when image searching on Internet or in large digital archives. The main aim of the study: usage of multi-step segmentation for near-duplicate image recognition. The methods used in the study: clustering of image pixels brightness is used for segmentation. The recurrent neural network is used for clustering. To estimate images similarity the authors have applied the cosine distance between rank distributions of clusters cardinality. The results: The authors suggested the search patterns based on the rank distributions of brightness clusters cardinality. The paper introduces the experimental results on the near-duplicate image recognition based on application of the suggested search patterns. It is shown that the use of a multi-step segmentation and rank distribution of the brightness clusters cardinality allows determining reliably the near-duplicate of the original image with a high degree of distortion on them, up to the radius of the Gaussian distortion equal 8 pixels. Such an approach also allows solving reliably the inverse problem of detecting the original image even in its fivefold reduced copy with radius Gaussian distortion on it to 8 pixels. |
first_indexed | 2024-03-13T07:49:18Z |
format | Article |
id | doaj.art-b0c80a497d3d420c94e7f2e383783d8f |
institution | Directory Open Access Journal |
issn | 2500-1019 2413-1830 |
language | Russian |
last_indexed | 2024-03-13T07:49:18Z |
publishDate | 2019-05-01 |
publisher | Tomsk Polytechnic University |
record_format | Article |
series | Известия Томского политехнического университета: Инжиниринг георесурсов |
spelling | doaj.art-b0c80a497d3d420c94e7f2e383783d8f2023-06-02T21:08:25ZrusTomsk Polytechnic UniversityИзвестия Томского политехнического университета: Инжиниринг георесурсов2500-10192413-18302019-05-013245Application of multi-step image segmentation for near-duplicate image recognitionVictor NemirovskiyAlexander StoyanovThe urgency of the paper is caused by the need to detect image near-duplicate in computer vision systems, as well as when image searching on Internet or in large digital archives. The main aim of the study: usage of multi-step segmentation for near-duplicate image recognition. The methods used in the study: clustering of image pixels brightness is used for segmentation. The recurrent neural network is used for clustering. To estimate images similarity the authors have applied the cosine distance between rank distributions of clusters cardinality. The results: The authors suggested the search patterns based on the rank distributions of brightness clusters cardinality. The paper introduces the experimental results on the near-duplicate image recognition based on application of the suggested search patterns. It is shown that the use of a multi-step segmentation and rank distribution of the brightness clusters cardinality allows determining reliably the near-duplicate of the original image with a high degree of distortion on them, up to the radius of the Gaussian distortion equal 8 pixels. Such an approach also allows solving reliably the inverse problem of detecting the original image even in its fivefold reduced copy with radius Gaussian distortion on it to 8 pixels.http://izvestiya.tpu.ru/archive/article/view/1371imagepixelpoint mappingrecurrent neural networkclusteringsegmentation |
spellingShingle | Victor Nemirovskiy Alexander Stoyanov Application of multi-step image segmentation for near-duplicate image recognition Известия Томского политехнического университета: Инжиниринг георесурсов image pixel point mapping recurrent neural network clustering segmentation |
title | Application of multi-step image segmentation for near-duplicate image recognition |
title_full | Application of multi-step image segmentation for near-duplicate image recognition |
title_fullStr | Application of multi-step image segmentation for near-duplicate image recognition |
title_full_unstemmed | Application of multi-step image segmentation for near-duplicate image recognition |
title_short | Application of multi-step image segmentation for near-duplicate image recognition |
title_sort | application of multi step image segmentation for near duplicate image recognition |
topic | image pixel point mapping recurrent neural network clustering segmentation |
url | http://izvestiya.tpu.ru/archive/article/view/1371 |
work_keys_str_mv | AT victornemirovskiy applicationofmultistepimagesegmentationfornearduplicateimagerecognition AT alexanderstoyanov applicationofmultistepimagesegmentationfornearduplicateimagerecognition |