What if your camera could fix what your microphone can't hear? That's the question behind CHEAR — and it just earned first place.
At Hack4Innovation Bicocca, organized by Fondazione UniversitĂ Bicocca and UniversitĂ degli Studi di Milano-Bicocca, the ISLab team won the Huawei Technologies Italia challenge, beating out three competitive teams with a focused and practical idea.
The insight was simple but powerful: while video technology has leaped forward in recent years, audio enhancement has struggled to keep pace. CHEAR bridges that gap. By watching lip movements and visual cues in real time, this neuroscience-inspired, multi-modal pipeline reconstructs and enhances degraded speech on the fly — whether you're a vlogger battling background noise, a remote worker on a choppy call, a live streamer in a loud venue, or a visually impaired user who relies on clear audio.
The team behind this win:
- Luigi Celona, tenure-track researcher at ISLab;
- Eduard I. Stan, RTDa researcher at ISLab;
- Andrea Corsico, freshly graduated in Master's in Data Science after his internship at ISLab;
- Eduardo Mosca, Master's in Data Science student currently interning at ISLab;
- Arian Emami, first-year Master's student in AI for Science and Technology.
Five people, one shared idea, a well-deserved trophy — and a €2,000 prize to go with it. 🏆