Chaotic Sparrow Search Algorithm with Deep Transfer Learning Enabled Breast Cancer Classification on...

Area of Science:

• Medical Imaging
• Computational Pathology
• Artificial Intelligence in Oncology

Background:

• Histopathological diagnosis is the gold standard for breast cancer detection but is time-consuming and subjective.
• Automated analysis of histopathological images is crucial to overcome diagnostic challenges and workload.
• Deep learning has shown significant promise in advancing breast cancer pathological image classification.

Purpose of the Study:

• To introduce a novel deep learning model for accurate breast cancer classification using histopathological images.
• To develop an automated system that overcomes the limitations of manual histopathological analysis.
• To enhance the precision and efficiency of breast cancer diagnosis through advanced computational methods.

Main Methods:

• A novel chaotic sparrow search algorithm with deep transfer learning-enabled breast cancer classification (CSSADTL-BCC) model was developed.
• Gaussian filtering (GF) was used for noise reduction, and MixNet for feature extraction.
• A stacked gated recurrent unit (SGRU) model classified images, with hyperparameters optimized by the chaotic sparrow search algorithm (CSSA).

Main Results:

• The CSSADTL-BCC model demonstrated superior performance in breast cancer classification on a benchmark dataset.
• The model achieved high accuracy, outperforming existing state-of-the-art approaches.
• Hyperparameter optimization of the SGRU model using CSSA proved effective for histopathological image analysis.

Conclusions:

• The CSSADTL-BCC model offers a precise and automated solution for breast cancer classification from histopathological images.
• This novel approach, utilizing hyperparameter-tuned SGRU, represents a significant advancement in computational pathology.
• The study highlights the potential of deep learning and optimization algorithms to improve cancer diagnosis accuracy and efficiency.