Over 500 million people in the world are affected by diabetes, and about one-third of them are going to develop foot ulcers sooner or later. The current system uses a four-week treatment and follow-up protocol to determine whether the ulcer is healing.
But if people are not careful about their hospital visits, the ulcer can get worse quietly, can result in a serious infection, or even result in amputation. Yet the standard of care still asks everyone to wait.
A team at RMIT University in Melbourne thinks that a first appointment is enough. They built TherMUL, a smartphone-based thermal imaging tool that tries to flag a stubborn wound on day one.
How does a phone predict healing?
The setup is pretty simple. TherMUL pairs a phone with a FLIR One thermal camera, a clip-on accessory that sells for roughly $250 to $300 in the US. A clinician holds it about 20 cm from the foot and takes one picture. Nothing touches the wound.
The software reads the texture of the heat pattern inside the wound instead of the average temperature. Patchy, uneven warmth can hint at inflammation or poor blood flow, and the model learned to link those patterns to wounds that aren’t healing or closing.
Does it actually work?
Partly, and the researchers are upfront about it. Across 68 ulcers from 56 patients at two Melbourne hospitals, the model correctly spotted about 76% of wounds that failed to heal within 12 weeks, but got only 53% of the healing ones right. On a separate batch of seven wounds it had never seen, it called five correctly.
Those are modest numbers, but nobody has managed this at the first visit before, and the team is clear about why that is the part worth paying attention to.
“To the authors’ knowledge, this is the first study that has demonstrated the feasibility of predicting non-healing DFU status at initial presentation using single-image texture analysis in a prospective longitudinal cohort, enabling more timely and targeted support for clinical decision-making in wound care, especially benefiting remote regions.”
Who actually benefits from this?
The obvious benefits are for people who live hours from a specialist foot clinic. The follow-up visits cost time, money, and energy, and the ones who skip appointments are at a high risk of seeing their foot ulcers turn into something serious. A cheap camera and a phone in a local clinic could tell someone which patients need escalating now rather than in a month.
The researchers sum up where the work stands.
“This study demonstrates the early feasibility of TherMUL, an AI-enabled smartphone-based thermal imaging device to assist clinicians in predicting 12-week diabetic foot ulcer healing outcomes at the first clinical presentation, with moderate discriminative performance using interpretable texture features derived from a single, non-contact image. Providing automated predictions of healing outcomes from baseline images has the potential to support clinical decision-making and facilitate earlier identification of wounds at risk of delayed healing.”
There is still a catch. All of this happened in controlled consultation rooms, and thermal readings drift with room temperature, humidity, and airflow, so nobody knows yet how the tool behaves in somebody’s living room. Larger and more varied trials come next, and TherMUL is meant to support a clinician’s judgement rather than replace it.

