Class-wise Performance Metrics for Lung Cancer Histopathology Image Classification

Nepolian Vailankanni's dataset, published on figshare in April 2026, contains performance metrics comparing a novel machine learning approach to a baseline model for lung cancer diagnosis from histopathological images. The 5.5 KB XLS file likely contains metrics such as balanced accuracy scores derived from experiments on three public datasets, including LungHist700 with 691 images across three categories. The proposed method achieved a 97% balanced accuracy, compared to 86% for the baseline Vision Graph Convolutional Neural Network.

Use Cases

• Benchmarking new classification models against the reported 97% balanced accuracy based on the described multi-technique fusion approach.
• Analyzing class-wise performance differences for normal, adenocarcinoma, and squamous cell carcinoma categories mentioned in the description.
• Comparing the efficacy of feature selection and fusion techniques (e.g., Minimum Redundancy Maximum Relevance) for medical image classification.

Strengths

• Performance metrics are validated on three publicly available datasets (LungHist700, LC25000, TCGA UT), suggesting robustness.
• The proposed method achieved a high reported balanced accuracy of 97%, a significant improvement over the 86% baseline.
• Dataset is small (5.5 KB) and in XLS format, facilitating quick inspection and analysis.

Limitations

• Row count and specific column definitions are unknown; field semantics must be inferred after download.
• The dataset contains derived performance metrics, not the original histopathological images or raw features.
• Scale is limited to 5.5 KB, indicating a very small summary table rather than a primary data source.

Provenance

Source

figshare, author Nepolian Vailankanni

Collection Method

Derived from experimental results comparing machine learning models on histopathological image datasets.

Time Range

Publication date is 2026; experimental timeframe is not specified.

Freshness

Last updated 2026-04-27 17:29:20; freshness should be verified.

Geography

null