#artificial intelligence
Sep 2026
RootQuantV2: Adapting a Vision Foundation Model for Root-Trait Regression from Minirhizotron Imagery
RootQuantV2 repurposes legacy numeric archives for high-throughput, automated root trait estimation, replacing RootQuant's CNN backbone with a self-supervised ViT and adapting a frozen DINOv3 ViT-L/16 with a hybrid parameter-efficient scheme.
Kinjalk Parth, Sebastian Varela, A. Leakey
· arXiv.org · 0 citations