The Woundtech Analytics team presented a research paper titled “A Deep Multi-Modal Method for Patient Wound Healing Assessment” in the Medical Imaging Meets NeurIPS workshop at the Neural Information Processing Systems (NeurIPS) 2019 Conference. They proposed a deep multi-modal method that could predict the risk of hospitalization by collectively using a patient’s wound variables and wound images. Working hand-in-hand with Woundtech’s National Vice President of Clinical Operations, Dr. Minghsun Liu, the team developed a transfer learning-based wound assessment solution that predicted wound variables from wound images and subsequently, their healing trajectories.
“We were very excited to see Woundtech at NeurIPS Conference and that our research paper was accepted,” said Shahid Mohammed, Woundtech’s Business Intelligence Manager. “We worked hard for six months to aggregate the research. It was great to present in front of so many researchers and professionals.”
Mohammed said that comorbidities – like delays in treatment – increased wound-care costs across the board. The research points to early detection leading to savings, fewer amputations, and fewer deaths. Untreated wounds lead to expensive hospitalization and developing a novel model would facilitate the early detection of complexities in the wound.
NeurIPS is an Artificial Intelligence conference where international AI research teams to present their research to their peers and experts. The annual conference fosters the exchange of research on neural information processing systems in biological, technological, mathematical, and theoretical aspects. The conference focuses on presentations and discussions of peer-reviewed novel research, along with talks by leaders in the field. Most of the papers came from reputed institutes like MIT, Stanford, Caltech, Google Brain, and Facebook. The conference is very popular among AI professionals and consequently sold 13,000 tickets in under 15 minutes online.