Purdue University’s College of Engineering is busy developing imaging tools that could transform everything from healthcare imaging to smart manufacturing and autonomous navigation. Under the guidance of Professor Qi Guo, the team has introduced CT-Bound and MetaHDR. CT-Bound is a fresh approach for detecting boundaries in noisy images, addressing the common challenge of capturing fine structural details in low-light settings using a blend of convolutional architectures and transformer models. This method delivers real‑time accuracy without the need for extra fine‑tuning.
MetaHDR takes a different route by using a versatile metasurface to capture high‑dynamic range images in one shot. This means no more juggling multiple exposures or dealing with motion blur, which can be a real headache in applications like autonomous vehicles or microscopic imaging. With their findings already making the rounds at IEEE and in Optics Express, the Purdue team is now focused on transforming these prototypes into practical, industry‑ready tools, with potential uses in areas such as depth estimation and agricultural phenotyping.
