Table 3 Classification Baselines Results (F1-score).
From: Packed Fruits and Vegetables Visual Classification and Segmentation Benchmark
Supervision | Model | Species (34 classes) | Varieties (65 classes) | ||||||
|---|---|---|---|---|---|---|---|---|---|
All | Packed | Not Packed | All Multiview | All | Packed | Not Packed | All Multiview | ||
Zero-Shot | CLIP | 0.43 | 0.28 | 0.58 | 0.58 | 0.28 | 0.2 | 0.36 | 0.4 |
BioCLIP | 0.3 | 0.12 | 0.48 | 0.46 | 0.1 | 0.04 | 0.15 | 0.17 | |
BioCLIP (taxons) | 0.39 | 0.19 | 0.6 | 0.51 | 0.24 | 0.09 | 0.4 | 0.36 | |
LLaVA1.5 | 0.58 | 0.45 | 0.73 | — | 0.24 | 0.17 | 0.3 | — | |
Linear Probing | CLIP | 0.95 | 0.91 | 0.99 | 0.96 | 0.96 | 0.94 | 0.98 | 0.96 |
BioCLIP | 0.88 | 0.78 | 0.97 | 0.92 | 0.88 | 0.82 | 0.95 | 0.9 | |
Supervised | ConvNext | 0.96 | 0.96 | 0.96 | 0.98 | 0.95 | 0.94 | 0.96 | 0.97 |
Supervised with non-packet only | ConvNext | 0.76 | 0.54 | 0.98 | 0.85 | 0.71 | 0.45 | 0.97 | 0.81 |
