Cheng‐Lin Liu

Chinese Academy of Sciences, Capital University of Physical Education and Sports

Papers

Total Citations

250

H-Index

About

Cheng-Lin Liu is a leading researcher whose work spans computer vision, scene text detection, and wearable robotics. His most impactful contribution is in arbitrary shape scene text detection, where he developed the adaptive text region representation method. This work, published in 2019, has garnered over 222 citations, reflecting its significance in enabling real-time text translation, automatic information entry, and assistive technologies for the blind. By addressing the challenge of detecting text in complex, non-horizontal layouts, Liu’s method has become a cornerstone for practical applications in robot sensing and autonomous systems. Beyond text detection, Liu has ventured into human performance augmentation, co-authoring a 2023 study on evaluating wearable lower limb exoskeletons. This work, with 11 citations, proposes integrated metrics beyond metabolic cost to benchmark exoskeleton effectiveness, advancing the field of wearable robotics. Most recently, in 2024, he contributed to weakly-supervised part segmentation with foundation models (WPS-SAM), pushing boundaries in efficient computer vision. Liu’s diverse portfolio—from foundational scene text techniques to cutting-edge robotics evaluation—demonstrates his ability to solve real-world challenges, making his research invaluable for students and engineers alike.