DeepHP: A Gastric Mucosa Histopathology Image Collection for Helicobacter pylori Infection Diagnosis

Precision medicine is benefiting from emerging deep learning-based applications, and one such application is computational histopathological analysis. However, a significant challenge arises from the scarcity of properly annotated training image datasets required for generating accurate classification and detection models. This scarcity primarily stems from human-related factors that hinder the acquisition of well-annotated data. To address this issue, the present study introduces DeepHP, a meticulously curated public collection of histopathological images. The DeepHP database encompasses a vast collection of 394,926 histopathological images, out of which 111k were annotated as Helicobacter pylori positive and 283k as Helicobacter pylori-negative.