EUVIMEDEUVIMED entdecken
Open Access · CC BY-NC-SA 4.0

Brain Tumor Classification Based on a Dual-Backbone ECA Fusion Network Model

Huatao LI

Medical Research · 15.06.2026 · 8 (2) · 19-27

DOI: 10.6913/mrhk.080203 ↗

Kurz gefasst

Fachartikel in der Zeitschrift Medical Research, online erschienen am 15.06.2026. Verfasst von Huatao LI. Der Volltext ist frei zugänglich (Open Access, CC BY-NC-SA 4.0).

AbstractBibliografische AngabenZitierenVerwandte Studien

Abstract (Originalsprache)

In recent years, with the gradual maturation of deep learning technologies and the steady improvement in computational power, computer-aided diagnostic techniques have been increasingly applied in clinical disease diagnosis. To effectively localize lesions, determine their categories, and provide reliable diagnostic support for clinicians, this paper proposes a dual-backbone ECA fusion network, which mainly consists of three components: a dual-backbone deep feature extraction module, an attention module, and a feature fusion and classification module. The research method of this paper employs two backbone networks, ResNet50 and EfficientNetB0, as two branches. After the feature extraction process, the ECA attention mechanism is adopted to enhance the feature extraction ability of the lesion area. Finally, through feature fusion, the classification of glioma, meningioma, pituitary tumor and normal control is achieved. Through comparative experiments, it was found that the proposed dual convolutional branch model structure can identify different types of brain tumors. On the test set, it achieved an accuracy of 93.28%, an AUC of 98.28%, and a precision of 94.12%, achieving good results. At the same time, in order to improve the interpretability of the model, Grad-CAM++ was used to show the regions of interest of the network, which coincided with the lesion area, further verifying the effectiveness of our dual-branch model.

Zitieren

APA

LI, H. (2026). Brain Tumor Classification Based on a Dual-Backbone ECA Fusion Network Model. Medical Research, 8(2), 19-27. https://doi.org/10.6913/mrhk.080203

Vancouver

LI H. Brain Tumor Classification Based on a Dual-Backbone ECA Fusion Network Model. Medical Research. 2026;8(2):19-27. doi:10.6913/mrhk.080203

BibTeX
@article{LI2026,
  author = {LI, Huatao},
  title = {Brain Tumor Classification Based on a Dual-Backbone ECA Fusion Network Model},
  journal = {Medical Research},
  year = {2026},
  volume = {8},
  number = {2},
  pages = {19--27},
  doi = {10.6913/mrhk.080203}
}
Kostenlos starten →