CRAT: Advanced transformer-based deep learning algorithms in OCT image classification

[Display omitted] •Class-Re-Attention Transformers to enhance the accuracy and speed of diagnosing retinal diseases from OCT images.•Re-attention module to improve feature extraction capability through aggregating features between different attention heads.•Class-re-attention module to make the clas...

Ausführliche Beschreibung

Gespeichert in:
Bibliographische Detailangaben
Veröffentlicht in:Biomedical signal processing and control Jg. 104; S. 107544
Hauptverfasser: Yang, Mingming, Du, Junhui, Lv, Ruichan
Format: Journal Article
Sprache:Englisch
Veröffentlicht: Elsevier Ltd 01.06.2025
Schlagworte:
Class attention
Optical coherence tomography
Re-attention
Retinal disease classification
Transformer
Transformer
Optical coherence tomography
Class attention
Re-attention
Retinal disease classification
ISSN:1746-8094
Online-Zugang:Volltext
Tags: Tag hinzufügen
Keine Tags, Fügen Sie den ersten Tag hinzu!
Beschreibung
Zusammenfassung:[Display omitted] •Class-Re-Attention Transformers to enhance the accuracy and speed of diagnosing retinal diseases from OCT images.•Re-attention module to improve feature extraction capability through aggregating features between different attention heads.•Class-re-attention module to make the class token focus on the task of class prediction. The primary retinal optical coherence tomography (OCT) images usually have speckle noise, which may lower the diagnostic accuracy. In this research, we developed a transformer-based deep learning algorithm named Class-Re-Attention Transformers (CRAT), which presented advanced performance to quickly and accurately predict possible retinal diseases and further pathological changes from easily accessible OCT images. In this context, a comprehensive collection of 109,371 retinal OCT images was curated. This collection encompasses 24,562 images indicative of AMD, 37,494 images representative of CNV, 11,598 images associated with DME, 8,896 images depicting drusen, and 26,821 images classified as normal. Among them, 190 images are used as the external test set, and they are from Xi ’an Ninth Hospital. CRA can enhance the learning of deep features and the integration of classification information through the synergy of Re-attention mechanism and attention-like layer. The Re-attention block helps mitigate the risk of Attention collapse, while the class-attention Layer enhances the classification performance by specifically handling the relationship between Class labels and features. This enhancement facilitates efficient diagnosis, leveraging the extracted features. In order to assess the performance of CRAT, the accuracy, precision and recall rate, specificity, and F1 score were used as the main index, which provide a comprehensive performance evaluation of the proposed algorithm. The results demonstrated that the average accuracy, average precision, average recall, average specificity and average F1 score of the five eye categories (AMD, CNV, DME, Drusen and Normal) perform well on the internal test dataset, which reached 94.40%, 94.42%, 94.39%, 98.60%, and 97.76%, respectively. And the results on the external test dataset are 97.33%, 96.33%, 97.08%, 99.17%, and 98.74%, respectively. CRA block can reduce the influence of image noise on diagnostic results. The proposed method can help ophthalmologists to quickly and accurately predict the likely occurrence of retinal diseases.
ISSN:1746-8094
DOI:10.1016/j.bspc.2025.107544