New Journal Publication in IEEE Tran. on Radar Systems

​CV-SincNet: Learning Complex Sinc Filters from Raw Radar Data for Computationally Efficient Human Motion Recognition
S. Biswas, C. O. Ayna, S.Z. Gurbuz, A.C. Gurbuz 
IEEE Transactions on Radar Systems, 2023

Our proposed approach, CV-SincNet, offers high classification accuracy with low computational latency while taking as input the raw, complex RF data samples. As a result, it can have important implications for applications requiring real-time RF data classification. Our results are shown over a 100 sign RF dataset of American Sign Language. For more, please check out our full manuscript here