LungCancer105K (LC105K): A Large-Scale Histopathological Image Dataset for Lung Cancer Diagnosis

This paper presents LungCancer105K (LC105K), a comprehensive dataset of 105,000 histopathological images of lung tissue. The dataset consists of three primary classes (Benign, Adenocarcinoma, Squamous Cell Carcinoma) with two additional subclasses, each containing 21,000 images. Derived from LC25000 and expanded using Augmentor package, LC105K facilitates development of accurate computer-aided diagnosis systems for lung cancer. Images measure 768x768 or 512x512 pixels in JPEG format. This dataset enables robust training and validation of deep learning models for lung cancer diagnosis, promoting early detection and improved treatment outcomes.