Octapoda.ehf has seen significant technological progress over the past two years, particularly in object tracking and detection. Data processing speed has increased thirtyfold, and detection accuracy has improved from 50% to 95%. More precise object detection using neural networks has allowed us to explore biological processes with greater accuracy and high-resolution imagery, a long-standing request from the research community. In the coming weeks, with the help of the HPC, we aim to meet the quality targets we’ve set for our software’s performance. A video demonstration will accompany this news release.
MarPar frumgerðin nálgast fullnaðarstig fyrir rannsóknir í haf- og líffræðirannsóknum
Tæknilegur framgangur hjá Octapoda.ehf hefur verið mikill síðustu tvö ár, sérstaklega á sviði leitunar (tracking) og hlutgreiningar (object detection). Hraði gagnavinnslu hefur aukist þrjátíufalt og nákvæmni í leitun hefur aukist úr 50% í 95%. Nákvæmari hlutgreining með tauganetum hefur gert okkur kleift að skoða líffræðileg ferli með meiri nákvæmni og í hárri
