Bounding Box Annotation for Localization

Use Case: Used primarily to identify and quantify particular cells of interest, such as tumor cells or infectious agents (immunohistochemical biomarkers such as Ki67, PD-L1), facilitating focused analyses and treatment planning. Technique: Applying simple rectangular annotations to precisely localize specific cellular formations or anomalies within a tissue slide.

Keypoint Annotation for Structural Markings in Cells

Use Case: Key in the study of cellular behaviours and pathology development, such as tracking the progression of metastatic cancer cells or other pathologies that alter cellular structures. Also essential in developing AI tools that standardize scoring criteria in several tumors, such as Gleason score. Technique: Pinpointing critical structural landmarks within cells or tissue architectures which are vital for morphological studies.

Polygon Annotation for Irregularly Shaped Cells

Use Case: Incredibly useful in complex cases where pathologies do not conform to standard shapes, such as in certain types of melanoma or irregular fibrosis patterns. It is also the basis in the development of AI search tools such as content-based image retrieval (CBIR) specially in Digital Pathology. Technique: Drawing detailed polygons around irregularly shaped pathological features to capture the precise dimensions and shape.