Improving the annotation process in computational pathology: from manual to semi-automatic approaches in digital nephropathology

The development of reliable artificial intelligence (AI) algorithms in pathology depends on solid ground truth provided by meticulous annotation of whole slide images (WSI), a time-consuming and operator-dependent process. A benchmark of the available annotation tools is performed to standardize and streamline this process.

在病理学领域可靠的人工智能（AI）算法开发，依赖于对全切片图像（WSI）进行细致入微标注所提供的坚实真实信息，这一过程耗时且受操作者依赖。对现有标注工具进行了基准测试，以实现标注过程的标准化和流程化。