The People's Daily Online Beijing Channel published an article on March 12, 2025, titled "First 'Spectral Convolutional Neural Network Chip' Proposed, Reducing Data Throughput by 96%" (https://www.peopleapp.com/column/30048490374-500006139614), which introduced the latest research achievements of our lab on spectral sensing chips. The report has garnered widespread attention, with page views exceeding 177,000 to date.

The advantages of our solution primarily lie in its high level of integration and low power consumption. It is capable of directly sensing natural light without relying on coherent light sources and can adaptively perform real-time calculations based on the imaging speed of CMOS image sensors. This enables the acquisition and processing of high-dimensional spectral images even on edge devices and mobile terminals with limited computational power. Our research group has long been dedicated to the field of on-chip spectral imaging, delving into how to empower terminal devices with spectral sensing and intelligent computing. The related achievements have also been industrialized, establishing a leading advantage in the integration of industry, academia, and research.