Melola Discovery

← Melola Discovery

Akses terbuka Jurnal SINTA 1

Hybrid dual-stream deep learning for breast cancer ultrasound detection

Bulletin of Electrical Engineering and Informatics · Vol. 15 No. 3 (2026) · hlm. 2467–2479

DOI
10.11591/eei.v15i3.10102
ISSN
2089-3191, 2302-9285
Terbit
2026-06-01
Penerbit
Institute of Advanced Engineering and Science (IAES)
Bahasa
en
Lisensi
creativecommons.org/licenses/by-sa/4.0/

Abstrak

The heterogeneity of breast tissue and subtle morphological variations in ultrasound images make breast cancer detection a challenging task. This study proposes a hybrid deep learning framework that integrates EfficientNetB4 and ConvNeXt within a dual-stream architecture enhanced by advanced attention mechanisms. The model combines multi-scale texture representation with spatial feature extraction to improve classification performance. A two-stage preprocessing pipeline, consisting of adaptive median filtering and bilateral filtering, is applied to reduce speckle noise while preserving important structural details. The proposed method is evaluated on BUSI and UDAIT datasets, achieving 87.82% accuracy, 87.33% precision, and 85.33% recall on BUSI, and 85.69% accuracy, 84.00% precision, and 78.00% recall on UDAIT. These results outperform several baseline models, including ResNet-50, DenseNet-121, and vision transformers. Error analysis shows limitations in detecting small lesions and cross-modal generalization, with reduced performance on mammography images. Attention visualization demonstrates strong agreement with radiologist annotations, supporting model interpretability. The findings highlight the effectiveness of hybrid architectures for ultrasound-based breast cancer detection while emphasizing the need for modality-specific optimization.

Kata kunci

Jaringan kolaborasi para penulis

40 penulis · 153 tautan kepenulisan bersama · 76 lintas organisasi · menampilkan 40 yang paling banyak terhubung

Arahkan kursor atau ketuk sebuah nama untuk menyorot rekan penulisnya. Klik dua kali untuk membuka halaman penulis.

  • Zarqa University
  • Al-Ahliyya Amman University
  • Benha University
  • Applied Science Private University
  • Al-Balqa Applied University
  • Al al-Bayt University
  • Hashemite University
  • Organisasi lain
  • Afiliasi tidak diketahui
  • Organisasi berbeda
  • Organisasi sama
  • Afiliasi tidak diketahui
  • Garis lebih tebal = lebih banyak artikel bersama · lingkaran lebih besar = lebih banyak artikel
Tampilkan sebagai tabel
Rekan penulisOrganisasiArtikel bersama
Hamza Abu OwidaAl-Ahliyya Amman University8
Suhaila AbuowaidaAfiliasi tidak diketahui8
Jamal Al-NabulsiAl-Ahliyya Amman University5
Nidal M. TurabAl-Ahliyya Amman University5
Muhammad Saleh Al-AyyadAl-Ahliyya Amman University4
Raed AlazaidahZarqa University4
Zaid MustafaAl-Balqa Applied University3
Nour Al HawamdehAl-Ahliyya Amman University2
Ahmad AburommanZarqa University2
Ali Mohd AliAl-Ahliyya Amman University1
Mohammad R. HassanAl-Ahliyya Amman University1
Feras AlnaimatAl-Ahliyya Amman University1
Ashraf Al SharahAl-Balqa Applied University1
Ghayth AlMahadinAl-Ahliyya Amman University1
Zaher SalahHashemite University1
Esraa Abu ElsoudZarqa University1
Waleed T. Al-SitHigher Colleges of Technology1
Esraa AlhenawiZarqa University1
Fuad Sameh AlshraiedehZarqa University1
Ghassan SamaraZarqa University1
Mohammad AljaidiZarqa University1
Hussain Mohammed TurkiZarqa University1
Essam Al-DaoudZarqa University1
Mais Haj QasemZarqa University1
Hani H. AttarZarqa University1
Ahmed Gamal Soliman Soliman DeabesModern University for Information and Technology1
Hala Abd El-kader MansourBenha University1
Michael Nasief MikhealBenha University1
Esraa M. EidBenha University1
Fadi Maher Al-KhasawnehKing Khalid University1
Ibrahim Fathi HuwariZarqa University1
Kamal Ahmad AlruzziZarqa University1
Luqman Mahmoud RababahJadara University1

Afiliasi berasal dari OpenAlex; metadata OAI-PMH jurnal tidak memuatnya. Penulis tanpa ID OpenAlex dicocokkan berdasarkan nama, sehingga dua orang bisa tergabung atau satu orang bisa terpisah.

Sumber metadata: OpenAlex API (jurnal SINTA, ADR 0048) https://openalex.org/S2765016208, pertama di-harvest 19 Nov 2026, terakhir terlihat 19 Nov 2026.
Metadata asli (oai_dc)
<oai_dc:dc xmlns:dc="http://purl.org/dc/elements/1.1/" xmlns:oai_dc="http://www.openarchives.org/OAI/2.0/oai_dc/"><dc:title xml:lang="en">Hybrid dual-stream deep learning for breast cancer ultrasound detection</dc:title><dc:creator>Musab Iqtait</dc:creator><dc:creator>Marwan Harb Alqaryouti</dc:creator><dc:creator>Ala Eddin Sadeq</dc:creator><dc:creator>Jafar Ababneh</dc:creator><dc:creator>Suhaila Farhan Ahmad Abuowaida</dc:creator><dc:creator>Nawaf Alshdaifat</dc:creator><dc:creator>Muath Alali</dc:creator><dc:subject>breast cancer detection</dc:subject><dc:subject>breast ultrasound</dc:subject><dc:subject>dual-stream architecture</dc:subject><dc:subject>EfficientNet-B4</dc:subject><dc:subject>ConvNeXt</dc:subject><dc:subject>attention mechanism</dc:subject><dc:subject>speckle reduction</dc:subject><dc:subject>adaptive median filter</dc:subject><dc:subject>bilateral filter</dc:subject><dc:subject>multi-scale texture features</dc:subject><dc:subject>BUSI dataset</dc:subject><dc:subject>small lesion detection</dc:subject><dc:subject>attention visualization</dc:subject><dc:subject>model interpretability</dc:subject><dc:description xml:lang="en">The heterogeneity of breast tissue and subtle morphological variations in ultrasound images make breast cancer detection a challenging task. This study proposes a hybrid deep learning framework that integrates EfficientNetB4 and ConvNeXt within a dual-stream architecture enhanced by advanced attention mechanisms. The model combines multi-scale texture representation with spatial feature extraction to improve classification performance. A two-stage preprocessing pipeline, consisting of adaptive median filtering and bilateral filtering, is applied to reduce speckle noise while preserving important structural details. The proposed method is evaluated on BUSI and UDAIT datasets, achieving 87.82% accuracy, 87.33% precision, and 85.33% recall on BUSI, and 85.69% accuracy, 84.00% precision, and 78.00% recall on UDAIT. These results outperform several baseline models, including ResNet-50, DenseNet-121, and vision transformers. Error analysis shows limitations in detecting small lesions and cross-modal generalization, with reduced performance on mammography images. Attention visualization demonstrates strong agreement with radiologist annotations, supporting model interpretability. The findings highlight the effectiveness of hybrid architectures for ultrasound-based breast cancer detection while emphasizing the need for modality-specific optimization.</dc:description><dc:publisher>Institute of Advanced Engineering and Science (IAES)</dc:publisher><dc:date>2026-06-01</dc:date><dc:type>info:eu-repo/semantics/article</dc:type><dc:format>application/pdf</dc:format><dc:identifier>https://doi.org/10.11591/eei.v15i3.10102</dc:identifier><dc:identifier>10.11591/eei.v15i3.10102</dc:identifier><dc:source>Bulletin of Electrical Engineering and Informatics; Vol 15 No 3 (2026); 2467-2479</dc:source><dc:source>2089-3191</dc:source><dc:source>2302-9285</dc:source><dc:language>en</dc:language><dc:relation>https://beei.org/index.php/EEI/article/download/10102/4789</dc:relation><dc:rights>https://creativecommons.org/licenses/by-sa/4.0/</dc:rights></oai_dc:dc>