Powered by OpenAIRE graph
Found an issue? Give us feedback
image/svg+xml art designer at PLoS, modified by Wikipedia users Nina, Beao, JakobVoss, and AnonMoos Open Access logo, converted into svg, designed by PLoS. This version with transparent background. http://commons.wikimedia.org/wiki/File:Open_Access_logo_PLoS_white.svg art designer at PLoS, modified by Wikipedia users Nina, Beao, JakobVoss, and AnonMoos http://www.plos.org/ Intelligent Systems ...arrow_drop_down
image/svg+xml art designer at PLoS, modified by Wikipedia users Nina, Beao, JakobVoss, and AnonMoos Open Access logo, converted into svg, designed by PLoS. This version with transparent background. http://commons.wikimedia.org/wiki/File:Open_Access_logo_PLoS_white.svg art designer at PLoS, modified by Wikipedia users Nina, Beao, JakobVoss, and AnonMoos http://www.plos.org/
Intelligent Systems with Applications
Article . 2024 . Peer-reviewed
License: CC BY
Data sources: Crossref
image/svg+xml art designer at PLoS, modified by Wikipedia users Nina, Beao, JakobVoss, and AnonMoos Open Access logo, converted into svg, designed by PLoS. This version with transparent background. http://commons.wikimedia.org/wiki/File:Open_Access_logo_PLoS_white.svg art designer at PLoS, modified by Wikipedia users Nina, Beao, JakobVoss, and AnonMoos http://www.plos.org/
https://dx.doi.org/10.60692/wx...
Other literature type . 2024
Data sources: Datacite
https://dx.doi.org/10.60692/53...
Other literature type . 2024
Data sources: Datacite
DBLP
Article
Data sources: DBLP
versions View all 5 versions
addClaim

This Research product is the result of merged Research products in OpenAIRE.

You have already added 0 works in your ORCID record related to the merged Research product.

Robust deep image clustering using convolutional autoencoder with separable discrete Krawtchouk and Hahn orthogonal moments

تجميع الصور العميقة القوية باستخدام الترميز التلقائي الملتف مع لحظات متعامدة منفصلة من Krawtchouk و Hahn
Authors: A. Bouali; Ilham El Ouariachi; Azeddine Zahi; Khalid Zenkouar;

Robust deep image clustering using convolutional autoencoder with separable discrete Krawtchouk and Hahn orthogonal moments

Abstract

En apprenant de manière coopérative les fonctionnalités et en attribuant des clusters, le clustering profond est supérieur aux algorithmes de clustering conventionnels. De nombreux algorithmes de clustering profond ont été développés pour une variété de niveaux d'application ; cependant, la majorité est toujours incapable d'apprendre des fonctionnalités latentes robustes et résistantes au bruit, ce qui limite les performances de clustering. Pour relever ce défi de recherche ouverte, nous introduisons, pour la première fois, une nouvelle approche appelée : Robust Deep Embedded Image Clustering algorithm with Separable Krawtchouk and Hahn Moments (RDEICSKHM). Notre approche tire parti des avantages des moments de Krawtchouk et de Hahn, tels que l'extraction de caractéristiques locales, l'orthogonalité discrète et la tolérance au bruit, pour obtenir une représentation d'image significative et robuste. De plus, nous utilisons LayerNormalization pour améliorer davantage la qualité de l'espace latent et faciliter le processus de clustering. Nous évaluons notre approche sur quatre ensembles de données d'image : MNIST, MNIST-test, USPS et Fashion-MNIST. Nous comparons notre méthode avec plusieurs méthodes de clustering approfondies basées sur deux métriques : la précision du clustering (ACC) et l'information mutuelle normalisée (NMI). Les résultats expérimentaux montrent que notre méthode atteint des performances supérieures ou compétitives sur tous les ensembles de données, démontrant son efficacité et sa robustesse pour le clustering d'images en profondeur.

Al aprender de forma cooperativa las características y asignar clústeres, el clúster profundo es superior a los algoritmos de clúster convencionales. Se han desarrollado numerosos algoritmos de clúster profundo para una variedad de niveles de aplicación; sin embargo, la mayoría sigue siendo incapaz de aprender características latentes robustas resistentes al ruido, lo que limita el rendimiento del clúster. Para abordar este desafío de investigación abierta, presentamos, por primera vez, un nuevo enfoque llamado: Robust Deep Embedded Image Clustering algorithm with Separable Krawtchouk and Hahn Moments (RDEICSKHM). Nuestro enfoque aprovecha las ventajas de los momentos Krawtchouk y Hahn, como la extracción de características locales, la ortogonalidad discreta y la tolerancia al ruido, para obtener una representación de imagen significativa y sólida. Además, empleamos LayerNormalization para mejorar aún más la calidad del espacio latente y facilitar el proceso de agrupamiento. Evaluamos nuestro enfoque en cuatro conjuntos de datos de imágenes: MNIST, MNIST-test, USPS y Fashion-MNIST. Comparamos nuestro método con varios métodos de agrupamiento profundo basados en dos métricas: precisión de agrupamiento (ACC) e información mutua normalizada (NMI). Los resultados experimentales muestran que nuestro método logra un rendimiento superior o competitivo en todos los conjuntos de datos, lo que demuestra su efectividad y robustez para el agrupamiento profundo de imágenes.

By cooperatively learning features and assigning clusters, deep clustering is superior to conventional clustering algorithms. Numerous deep clustering algorithms have been developed for a variety of application levels; however, the majority are still incapable of learning robust noise-resistant latent features, which limits the clustering performance. To address this open research challenge, we introduce, for the first time, a new approach called: Robust Deep Embedded Image Clustering algorithm with Separable Krawtchouk and Hahn Moments (RDEICSKHM). Our approach leverages the advantages of Krawtchouk and Hahn moments, such as local feature extraction, discrete orthogonality, and noise tolerance, to obtain a meaningful and robust image representation. Moreover, we employ LayerNormalization to further improve the latent space quality and facilitate the clustering process. We evaluate our approach on four image datasets: MNIST, MNIST-test, USPS, and Fashion-MNIST. We compare our method with several deep clustering methods based on two metrics: clustering accuracy (ACC) and normalized mutual information (NMI). The experimental results show that our method achieves superior or competitive performance on all datasets, demonstrating its effectiveness and robustness for deep image clustering.

من خلال ميزات التعلم التعاوني وتعيين المجموعات، يتفوق التجميع العميق على خوارزميات التجميع التقليدية. تم تطوير العديد من خوارزميات التجميع العميق لمجموعة متنوعة من مستويات التطبيق ؛ ومع ذلك، لا تزال الغالبية غير قادرة على تعلم ميزات كامنة قوية مقاومة للضوضاء، مما يحد من أداء التجميع. لمواجهة هذا التحدي البحثي المفتوح، نقدم، لأول مرة، نهجًا جديدًا يسمى: خوارزمية قوية لتجميع الصور المضمنة العميقة مع لحظات Krawtchouk و Hahn المنفصلة (RDEICSKHM). يستفيد نهجنا من مزايا لحظات كراوتشوك وهان، مثل استخراج السمات المحلية، والتعامد المنفصل، وتحمل الضوضاء، للحصول على تمثيل صورة هادف وقوي. علاوة على ذلك، نستخدم LayerNormalization لزيادة تحسين جودة المساحة الكامنة وتسهيل عملية التجميع. نقوم بتقييم نهجنا على أربع مجموعات بيانات للصور: MNIST واختبار MNIST و USPS و Fashion - MNIST. نقارن طريقتنا بالعديد من طرق التجميع العميق بناءً على مقياسين: دقة التجميع (ACC) والمعلومات المتبادلة الطبيعية (NMI). تُظهر النتائج التجريبية أن طريقتنا تحقق أداءً متفوقًا أو تنافسيًا في جميع مجموعات البيانات، مما يدل على فعاليتها وقوتها في تجميع الصور العميق.

Keywords

Artificial intelligence, Orthogonal polynomials, Convolutional neural network, Image Retrieval, Separable space, Pattern recognition (psychology), Mathematical analysis, Hahn moments, Cluster analysis, Discrete orthogonal polynomials, Image Feature Retrieval and Recognition Techniques, Shape Matching and Object Recognition, FOS: Mathematics, Deep clustering, Discrete separable orthogonal moments, Feature Selection, Wilson polynomials, Spectral Clustering, Interest Point Detectors, Deep learning, QA75.5-76.95, Autoencoder, Kravchuk polynomials, Image clustering, Computer science, Algorithm, Krawtchouk moments, Combinatorics, Electronic computers. Computer science, Computer Science, Physical Sciences, Q300-390, Computer Vision and Pattern Recognition, Content-Based Image Retrieval, Face Recognition and Dimensionality Reduction Techniques, Cybernetics, Mathematics

  • BIP!
    Impact byBIP!
    selected citations
    These citations are derived from selected sources.
    This is an alternative to the "Influence" indicator, which also reflects the overall/total impact of an article in the research community at large, based on the underlying citation network (diachronically).
    2
    popularity
    This indicator reflects the "current" impact/attention (the "hype") of an article in the research community at large, based on the underlying citation network.
    Top 10%
    influence
    This indicator reflects the overall/total impact of an article in the research community at large, based on the underlying citation network (diachronically).
    Average
    impulse
    This indicator reflects the initial momentum of an article directly after its publication, based on the underlying citation network.
    Average
Powered by OpenAIRE graph
Found an issue? Give us feedback
selected citations
These citations are derived from selected sources.
This is an alternative to the "Influence" indicator, which also reflects the overall/total impact of an article in the research community at large, based on the underlying citation network (diachronically).
BIP!Citations provided by BIP!
popularity
This indicator reflects the "current" impact/attention (the "hype") of an article in the research community at large, based on the underlying citation network.
BIP!Popularity provided by BIP!
influence
This indicator reflects the overall/total impact of an article in the research community at large, based on the underlying citation network (diachronically).
BIP!Influence provided by BIP!
impulse
This indicator reflects the initial momentum of an article directly after its publication, based on the underlying citation network.
BIP!Impulse provided by BIP!
2
Top 10%
Average
Average
gold
Related to Research communities