Home  |  Team  |  Research  |  Publications  |  Projects

KiNet: A Computerized Method for Ki67 Labeling Index Assessment

The central goal of this project is to develop a computerized method, namely KiNet, for Ki67 labeling (LI) index assessment in Ki67 immohistochemistry stained images of gastrointestinal and pancreatic neuroendocrine tumors (NETs). We use deep neural networks to implement an end-to-end learning method to identify different types of nuclei (i.e., immunopositive tumor, immunonegative tumor and non-tumor) for Ki67 LI measurement.

Source codes: [KiNet_v1]

Adversarial Domain Adaptation and Pseudo-Labeling for Cross-Modality Microscopy Image Quantification
Xing et al., The 22nd International Conference on Medical Image Computing and Computer Assisted Intervention (MICCAI), vol. 11764, pp.740-749, 2019

Generative Adversarial Domain Adaptation for Nucleus Quantification in Images of Tissue Immunohistochemically Stained for Ki-67
Zhang et al., JCO Clinical Cancer Informatics, vol. 4, pp.666-679, 2020 [Codes]

Bidirectional Mapping-Based Domain Adaptation for Nucleus Detection in Cross-Modality Microscopy Images
Xing et al., accepted to IEEE Transactions on Medical Imaging, 2020

Home  |  Team  |  Research  |  Publications  |  Projects