This service automatically analyzes colonoscopy video recordings to calculate withdrawal time per anatomical colon segment, a key quality indicator of the procedure. The system identifies colonic regions frame by frame, separates useful inspection time from non-informative intervals (cleaning, close contact with the colon wall), and generates a detailed report with quality indicators comparable to clinical guideline thresholds. The main goal is to support gastroenterologists in self-evaluating and continuously improving their withdrawal technique, through objective and reproducible measurements independent of observer subjectivity.

This toolkit is designed to automatically analyze functional brain connectivity based on resting-state fMRI (rs-fMRI) data and to identify deviations from a model considered 'normal,' highlighting changes associated with psychiatric disorders and chronic pain. In practice, the system constructs a reference model of how brain networks interact in healthy individuals and subsequently compares patient data against this standard to detect abnormal connectivity patterns. The primary goal is to support the discovery of objective neuroimaging biomarkers, which could contribute to more precise diagnosis and personalized medicine.

This service provides a deep learning pipeline for the automated morphological segmentation of colorectal cancer within Whole Slide Images (WSI). The model performs pixel-level segmentation into five distinct histological categories — stroma, adenocarcinoma invasion, high-grade and low-grade dysplasia, and normal gland — reducing inter-observer variability and supporting pathologists and gastroenterologists in efficiently identifying critical histological regions of interest.