Accuracy Comparison of Related Works for Lung Cancer Histopathology Image Classification

691 histopathological images extracted from the publicly available LungHist700 dataset, categorized into 151 normal subjects, 280 lung adenocarcinoma subjects, and 260 lung squamous cell carcinoma subjects. The dataset, authored by Nepolian Vailankanni and last updated in April 2026, contains results from a study proposing a novel deep learning approach for lung cancer diagnostics. The proposed method achieved a reported balanced accuracy of 97% on this data.

Use Cases

• Benchmarking new lung cancer classification models based on reported accuracy metrics.
• Analyzing the performance impact of feature fusion and selection techniques described in the research.
• Comparing the efficacy of different self-supervised learning methods (Deep Cluster, BYOL, SimCLR) for medical image analysis.

Strengths

• Provides a direct accuracy comparison (97% vs 86%) for a proposed novel method against a baseline model.
• Results are validated across three publicly available datasets (LungHist700, LC25000, TCGA UT).
• Data is openly shared under a CC-BY-4.0 license.

Limitations

• The dataset is very small at 5.5 KB; the underlying image data is not included, only the comparison results.
• Column-level documentation is absent; field semantics must be inferred after download.
• Row count is unknown, which may limit suitability assessment.

Provenance

Source

figshare

Collection Method

Images extracted from the LungHist700 dataset; results from a research study applying deep learning techniques.

Time Range

null

Freshness

Last updated 2026-04-27 17:29:24; freshness should be verified.

Geography

null