Marcelo Mendonça Luciano Oliveira Memorizing Features Efficiently for Self-supervised Video Object Segmentation. 2026 443-452 32 J. Braz. Comput. Soc. 1 https://doi.org/10.5753/jbcs.2026.5904 db/journals/jbcs/jbcs32.html#MendoncaO26 streams/journals/jbcs