vividRGB: Various visual cue detection in RGB photographs with forced focused attention
Organizations: Leibniz Centre for Agricultural Landscape Research (ZALF), Eberswalder Straße 84, 15374 Müncheberg, Germany
Abstract
Natural photographs contain dense, overlapping structures that can be weakly expressed in feature maps from whole-image vision encoders. We introduce forced focused attention (FFA) and its Python implementation, vividRGB. FFA is an encoder-agnostic inference-time procedure that applies a frozen encoder to enlarged, overlapping local crops, averages inverse-aligned reflections, and reconstructs a feature field in source coordinates. We evaluate 50 high-resolution photographs from grassland vegetation, fungal networks, airborne forests, coral reefs, and phytoplankton microscopy, using DINOv3 ViT-B/16. Across the five domains, FFA at 5% or 10% focus produces higher mean RGB-edge boundary F1 than direct DINOv3 encoding and automatic SAM3 masks. Domain means range from 0.244 to 0.596 for FFA, versus 0.028-0.076 for direct DINOv3 and 0.020-0.090 for SAM3. Qualitative feature maps also show finer appearance structure. These results concern RGB-edge preservation rather than verified species, traits, or biological segmentation. Narrower focus generally improves boundary preservation, while extraction time and GPU energy use increase sharply and gains diminish at finer settings. FFA may be useful for relatively small, high-resolution ecological image collections where dense visual cues matter. The alpha release vividRGB 0.1.0 and versioned research resources support reproducible cue inspection.
Figures & tables
| Domain | Source and content | Working dimensions (pixels) |
|---|---|---|
| Grassland | MOIN/KICS-Zert; mixed vegetation at four German sites | |
| Fungal networks | Zenodo 5725751; branching mycelial networks [ 9 , 36 ] | |
| Airborne forest | NeonTreeEvaluation; NEON RGB canopy imagery [ 10 ] | to , including nonsquare windows |
| Coral reefs | Coralscapes; benthic community frames [ 8 ] | |
| Phytoplankton | PMID2019; laboratory microscopy [ 11 ] | or |
| Domain | DINOv3 original | SAM3 automatic | DINOv3-FFA 5% | DINOv3-FFA 10% |
|---|---|---|---|---|
| Grassland (MOIN) | 0.061 0.026 | 0.020 0.016 | 0.596 0.106 | 0.441 0.083 |
| Fungal networks | 0.076 0.057 | 0.034 0.027 | 0.290 0.108 | 0.244 0.093 |
| Airborne forest (NEON) | 0.034 0.011 | 0.041 0.044 | 0.549 0.136 | 0.360 0.135 |
| Coral reefs (Coralscapes) | 0.041 0.010 | 0.036 0.024 | 0.436 0.136 | 0.266 0.074 |
| Phytoplankton (PMID2019) | 0.028 0.007 | 0.090 0.032 | 0.456 0.118 | 0.288 0.138 |
| Method | Duration (s) | Board energy (kJ) | Efficiency ( ) |
|---|---|---|---|
| SAM3 automatic | 5.0 0.3 | 0.32 0.02 | 3.104 0.181 |
| DINOv3 original | 0.317 0.165 | 0.0210 0.0109 | 70.072 46.793 |
| DINOv3-FFA100% | 0.328 0.060 | 0.0093 0.0017 | 111.835 26.628 |
| DINOv3-FFA40% | 3.1 0.7 | 0.15 0.05 | 7.053 1.803 |
| DINOv3-FFA20% | 11.4 2.8 | 0.72 0.21 | 1.480 0.318 |
| DINOv3-FFA10% | 43.8 11.9 | 2.96 0.83 | 0.3585 0.0764 |
Appendix figures & tables7 assets
Supplementary material from the paper’s appendix.
Appendix
| Experiment or control | Recorded settings |
|---|---|
| Primary DINOv3 comparison | Native single-orientation control; FFA 5% and 10%; 384-pixel input, 75% overlap, four reflections; six-treatment shared PCA16; independent (12 sensitivity). |
| Reconstruction | Crop minimum 16; bilinear placement; per-axis Hann floor 0.05; positive coverage normalization; no token L2 normalization or feature whitening. |
| Automatic SAM3 | 1008-pixel input; internal point grid; batch 4; no crop layers; IoU 0.88, stability 0.95, NMS 0.7; highest-IoU overlap priority. |
| RGB-edge endpoint | Gamma-encoded grayscale; Gaussian ; Sobel/8; threshold ; two-pixel erosion and Manhattan tolerance. |
| Automatic grouping | Long-side-256 grids; PCA ; K-means –12; full-covariance GMM –12; post hoc ; sensitivity –64. |
| Crop-size analysis | Separate calibration; eight complete 50-image cohorts, 100%–2%; four reflections; two reconstruction branches in measured DINOv3 cost. |
| Domain | Focus | Baseline | 95% interval | Wins | Raw | Holm | |
|---|---|---|---|---|---|---|---|
| Grassland | 5% | DINOv3 | 0.535 | [0.460, 0.604] | 10/10 | 0.00195 | 0.03906 |
| Grassland | 5% | SAM3 | 0.575 | [0.505, 0.642] | 10/10 | 0.00195 | 0.03906 |
| Grassland | 10% | DINOv3 | 0.380 | [0.326, 0.434] | 10/10 | 0.00195 | 0.03906 |
| Grassland | 10% | SAM3 | 0.421 | [0.367, 0.476] | 10/10 | 0.00195 | 0.03906 |
| Fungi | 5% | DINOv3 | 0.214 | [0.144, 0.289] | 10/10 | 0.00195 | 0.03906 |
| Fungi | 5% | SAM3 | 0.257 | [0.181, 0.327] | 10/10 | 0.00195 | 0.03906 |
| DINOv3-FFA5% | DINOv3-FFA10% | |||||||
|---|---|---|---|---|---|---|---|---|
| Mean | Median | –6 | Cap | Mean | Median | –6 | Cap | |
| 1 | 11.14 | 11 | 0% | 48% | 11.90 | 12 | 0% | 96% |
| 2 | 8.10 | 8 | 14% | 8% | 10.26 | 11 | 4% | 36% |
| 4 | 5.34 | 5 | 74% | 0% | 6.34 | 6 | 64% | 0% |
| 8 | 3.38 | 3 | 34% | 0% | 4.04 | 4 | 56% | 0% |
| 16 | 2.38 | 2 | 8% | 0% | 2.64 | 3 | 10% | 0% |
| Features | Mean | Median | –6 | ||
|---|---|---|---|---|---|
| 3 | DINOv3 original | 10.02 | 11 | 10% | 42% |
| 3 | DINOv3-FFA5% | 5.60 | 5.5 | 60% | 0% |
| 3 | DINOv3-FFA10% | 6.58 | 6 | 46% | 4% |
| 16 | DINOv3 original | 11.90 | 12 | 0% | 92% |
| 16 | DINOv3-FFA5% | 5.34 | 5 | 74% | 0% |
| 16 | DINOv3-FFA10% | 6.34 | 6 | 64% | 0% |
| Range | Coordinate | Knot | Slope change | Frequency |
|---|---|---|---|---|
| 100%–2% | mean energy | 20% | Acceleration | 100.0% |
| 100%–2% | inverse focus | 20% | Acceleration | 100.0% |
| 100%–2% | Linear mean energy | 5% | Deceleration | 62.3% |
| 40%–2% | mean energy | 4% | Deceleration | 99.3% |
| 20%–2% | mean energy | 4% | Deceleration | 93.0% |
| DINOv3 | SAM3 image encoder | |||||
|---|---|---|---|---|---|---|
| Domain | Original | FFA10% | FFA5% | Original | FFA10% | FFA5% |
| Grassland | 0.0915 | 0.4837 | 0.5787 | 0.2487 | 0.4733 | 0.4799 |
| Fungal networks | 0.0314 | 0.2172 | 0.3346 | 0.1258 | 0.3040 | 0.3439 |
| Airborne forest | 0.0317 | 0.3979 | 0.6096 | 0.2997 | 0.5789 | 0.6062 |
| Coral reefs | 0.0511 | 0.2172 | 0.4144 | 0.2512 | 0.4522 | 0.4923 |
| Phytoplankton | 0.0240 | 0.0654 | 0.1637 | 0.1316 | 0.4004 | 0.4496 |
| Method | Precision | Recall | F1 | Density | 3-axis F1 |
|---|---|---|---|---|---|
| DINOv3 original | 0.4853 | 0.0244 | 0.0459 | 0.0040 | 0.1792 |
| DINOv3-FFA10% | 0.4368 | 0.2184 | 0.2763 | 0.0420 | 0.2864 |
| DINOv3-FFA5% | 0.4764 | 0.3984 | 0.4202 | 0.0761 | 0.3843 |
| SAM3 image encoder | 0.3403 | 0.1559 | 0.2114 | 0.0445 | 0.2654 |
| SAM3-FFA10% | 0.3083 | 0.8012 | 0.4417 | 0.4064 | 0.5317 |
| SAM3-FFA5% | 0.3227 | 0.9230 | 0.4744 | 0.5895 | 0.5284 |
Fine-Grained Visual Perception
Fine-grained visual perception studies how models detect, distinguish, and ground subtle visual details, attributes, parts, and changes.
See all 164 papers