Efficient robot vision compression using learned and standard formats

Original: SEAOTTER: Sensor Embedded Autoencoding with One-Time Transcode for Efficient Reconstruction

Source: arxiv.org ↗