Related Experiment Video
Updated: Feb 17, 2026
04:48
Application of Deep Learning-Based Medical Image Segmentation via Orbital Computed Tomography
Published on: November 30, 2022
3.6K
Automated Classification of Kidney Tumours Using Deep Convolutional Neural Networks
Miaolin Guo1, Shan Lin2, Yuxuan You3
1The School of Clinical Medicine, Fujian Medical University, Fuzhou, China.
Summary
This study introduces a novel deep learning model for kidney tumour detection using CT scans. The model achieves high accuracy in classifying renal tumours, improving early diagnosis efficiency.
Area of Science:
- Medical Imaging
- Artificial Intelligence
- Oncology
Background:
- Kidney tumour (KT) is a significant global health concern, ranking as the seventh most common cancer.
- Early detection of kidney tumours is vital for successful treatment outcomes.
- Current detection methods, like computed tomography (CT), are time-consuming and labor-intensive.
Purpose of the Study:
- To develop an efficient deep learning model for detecting and classifying kidney tumours from CT images.
- To enhance the accuracy and speed of renal tumour diagnosis.
Main Methods:
- Utilized dual deep learning backbones (EfficientNetV2-B3 and ResNet50) for robust feature extraction.
- Employed a dual efficient channel attention module (DECA) for effective cross-model feature fusion.
- Evaluated the model on the CT Kidney dataset.
Main Results:
- The proposed model achieved 98.80% accuracy and a 98.28% F1 score.
- Demonstrated superior performance compared to alternative feature fusion techniques such as cross-attention and weighted-sum.
Conclusions:
- The integrated dual backbone and DECA model shows significant promise for classifying renal tumour CT images.
- This AI-driven approach offers a powerful and efficient solution for kidney tumour diagnosis.
