How can I download the PathMiner software?
For security reasons, we require a login and password to download the PathMoner software. You can obtain a login name by emailing the administrator (link sends e-mail). Please include a brief description of who you are and your interests in the software.
What does the PathMiner software do?
The PathMiner algorithms and software were developed to provide the means for conducting interactive telemicroscopy while providing reliable decision support for the detection and diagnosis of lymphoproliferative disorders. The primary focus of the studies conducted thus far included a subset of malignant lymphomas and leukemias which are often confused with one another during routine microscopic evaluation. The PathMiner system features both single cell and multi-cell composite analysis (Please see publications and demonstration videos for further detail).
The image archive and correlated relational database provides the following features:
- Database Review: the user can view all the images and the correlated immunophenotype profiles of each case in the ground-truth database.
- Database Queries: the user can query the database using different diagnostic criteria including the diagnosis of record, the morphometric signature, or the immunophenotype.
- Image Segmentation Based on Robust Color GVF Snake: The automated indexing module of the PathMiner system utilizes a newly developed Robust Color GVF Snake (ref: http://pleiad.rwjms.rutgers.edu/pathminer ) Applications: Lin Yang, Peter Meer, David J. Foran. "Unsupervised Analysis of Imaged Pathology Specimens Based on Robust Estimation and Color Active Contour Models". IEEE Trans. on Information Technology in Biomedicine, Vol. 9, pp. 475-486, 2005.) to delineate both the nuclei and the cytoplasm of each imaged cell automatically.
- Image Indexing: Authorized users can distributed ground-truth databases with new cases.
- Rank Retrieval: Based upon a set of texture descriptors and thearea and shape signature of the nuclei and cytoplasm, PathMiner retrieves cases which have similar spectral and spatial signatures to a given unknown (unclassified) cell and provides the individual seeking decision support with a set of ranked retrieval results including imaged cells. Based upon the majority logic of those retrievals the system presents the user with the statistically most likely diagnosis.