TORUS

About us

Torus Actions SAS is an artificial intelligence company created in 2019 in Toulouse, France, by a group of scientists led by Nguyen Tien Zung, distinguished professor of mathematics. During the past two years, the company has grown very quickly, from a handful of people to more than thirty collaborators, including 20 PhDs in Mathematics and AI. The company develops many new AI algorithms, especially for image processing (satellite image processing, object recognition in videos, X-Ray/MRI/CT medical image analysis, etc.).

In particular, the SkinCancerAI program developed by Torus Actions can analyse dermatoscopic/clinical images of  skin lesions in a very detailed way and classify them into more than 100 different benign and malignant classes and subclasses, e.g., acral melanoma in situ, aggressive basal cell carcinoma, Kaposi sarcoma, eccrine poroma, sebaceous gland hyperplasia, Spitz nevus, etc. Professor Jean-Luc Perrot, chief dermatologist at CHU Saint-Etiennes, France, is co-founder and chief medical advisor of Torus Actions.

Torus Actions  won grants from the regional and national government of France, together with the "Deep Tech" label. The company already has some big-name pharmaceutical companies and governmental organizations as its partners and customers. It also as research collaboration with many universities and research institutes, including Harvard Medical School..

Learn more about us at www.torus.ai.

Main role in the project

The main role of Torus in the project is to develop the AI used in the triage, monitoring, image-based diagnosis, and personalized multi-modal diagnosis of the patients. In particular, our AI solutions will:

  • Match different images of the same lesion and monitor its evolution.
  • Give a triage and diagnosis of all skin lesions based on the images taken by the ITOBOS 3D scanner.
  • Explain the given diagnosis via visual features on the images (part of explainable AI).
  • Combine image analysis with metadata analysis to arrive at a more precise personalized diagnosis.