Company News

Panasonic R&D Center Singapore’s paper accepted for Publication in an Upcoming Issue of the Transactions on Pattern Analysis and Machine Intelligence (TPAMI)

September 24, 2020 | Project

Teaming up with Nanjing University of Science and Technology (NUST), Panasonic R&D Center Singapore co-authored a journal paper titled “Covariance Attention for Semantic Segmentation” which has been accepted for publication in an upcoming issue of the Transactions on Pattern Analysis and Machine Intelligence, the world’s top journal in the field of computer vision and pattern recognition.  In the article, we present a low complexity approach that exploits the covariance matrix to encode the dependencies over local and global cues of the scene.  The spatial and channel covariance attention modules are designed, respectively, to boost the accuracy of semantic segmentation.  We build a new deep learning network named CANet for scene parsing/semantic segmentation by using the proposed covariance attention modules to achieve the state-of-the-art performance on multiple challenging datasets.