Infoblock alpha
med/imaging IB-arxiv-2607.28148-f1-11101-samples measured

Under 5-fold cross-validation on a merged 11,101-sample TCM tongue-image dataset, the best model reached weighted-F1 of 0.7761.

Active Evidence state is recorded explicitly and may be revised by a later event.

Observed quantities

weighted_f1 0.7761 · n_samples 11101 · cv_folds 5

01

Primary source

  1. arxiv · 2607.28148 · v1 · abstract
Title
What Makes Deep Learning Work for Traditional Chinese Medicine Tongue Diagnosis? A Comprehensive Ablation Study
Authors
Longxia Gao, Linan Wang, Yuhe Han, Junze Geng, Meng Zhang, Hanqing Zhao
Version
v1
Check
Support quote found word for word in the arXiv abstract v1, 2026-10-08
Supporting quote from the source
the best 11,101-sample model reached weighted-F1 of 0.7761

02

Boundary of this block

No additional public limitation was recorded.

03

Connections

Outgoing

No outgoing public edges.

Incoming

No incoming public edges.