Fine-grained Image Classification on CU-Tree102 benchmark leaderboard · SOTA2 Research

Fine-grained Image Classification on CU-Tree102

86.65AccuracyCGL + Ours

Evaluation Results

Method	Links
CGL + Ours2026.01		86.65	79.83	77.35	77.87	86.31	86.65	86.17
HERBS + Ours2026.01		85.61	78.29	76.11	76.44	85.22	85.61	85.06
Swin-Base + Ours2026.01		83.44	75.83	72.86	73.36	82.78	83.44	82.68
MPSA + Ours2026.01		83.31	70.44	74.45	70.41	87	83.31	84.67
HERBS2026.01		82.98	74.05	70.91	71.26	81.98	82.98	81.9
Expert-FT2026.01		81.61	74.43	70.2	70.93	81.28	81.61	80.87
CGL2026.01		81.47	72.14	70.37	70.34	81.32	81.47	81.03
MPSA2026.01		80.84	65.07	65.5	63.8	87.21	80.84	83.32
Swin-Base2026.01		79.33	69.62	67.72	67.78	78.76	79.33	78.64
ViT-Base + Ours2026.01		78.07	69.83	65.81	66.6	77.4	78.07	77.2
ResNet-50 + Ours2026.01		77.96	69.81	66.73	67.25	77.88	77.96	77.42
ViT-Base2026.01		73.3	64.11	60.67	61.15	73.11	73.3	72.65
ResNet-502026.01		68.26	55.28	53.43	53.33	66.99	68.26	66.98