Caption Quality Test for Cross-View Geo-Localization

Validate whether generated text captions improve retrieval R@1 in cross-view geo-localization (drone-to-satellite).

Architecture

Overview

Asymmetric dual-encoder with domain-specific frozen backbones and hierarchical text fusion for drone-to-satellite image retrieval.

┌──────────────────────────── QUERY BRANCH ────────────────────────────┐
│                                                                       │
│  drone_img ──► DINOv3 ViT-L/16 LVD ──► CLS token                    │
│  [B,3,256,256]   (frozen, 303M)         d_img [B,1024]               │
│                                            │                          │
│  L1 (overview)    ──► DGTRS-CLIP ──► z₁ [B,768] ─┐                   │
│  L2 (full desc)   ──► DGTRS-CLIP ──► z₂ [B,768] ─┼─ cat ──► [B,2304]│
│  L3 (fingerprint) ──► DGTRS-CLIP ──► z₃ [B,768] ─┘    │             │
│      (248 tokens, KPS pos. emb.)          MLP(2304→1024→1024)        │
│                                                 │                     │
│                                            d_txt [B,1024]            │
│                                                 │                     │
│           q = σ(α_q)·d_img + (1−σ(α_q))·d_txt   GatedFusion_q      │
│                          │                                            │
│                    q̂ = q / ‖q‖₂  ──► query [B,1024]                  │
└───────────────────────────────────────────────────────────────────────┘

┌──────────────────────────── GALLERY BRANCH ──────────────────────────┐
│                                                                       │
│  sat_img ──► DINOv3 ViT-L/16 SAT ──► CLS token                      │
│  [B,3,256,256]  (frozen, 303M)        s_img [B,1024]                 │
│                                          │                            │
│  sat_L1 ──► DGTRS-CLIP ──► z₁ [768] ─┐                               │
│  sat_L2 ──► DGTRS-CLIP ──► z₂ [768] ─┼─ cat ──► MLP ──► s_txt [1024]│
│  sat_L3 ──► DGTRS-CLIP ──► z₃ [768] ─┘      (shared MLP)            │
│                                          │                            │
│       g = σ(α_g)·s_img + (1−σ(α_g))·s_txt   GatedFusion_g           │
│                                          │                            │
│                                   ĝ = g / ‖g‖₂  ──► gallery [B,1024]│
└───────────────────────────────────────────────────────────────────────┘

Retrieval space: 1024-dim (DINOv3 native, no projection layers)
TextFusionMLP shared between query and gallery branches
For sat images without captions: s_txt=None → g = s_img (gate passthrough)

BASELINE: σ(α) = 1.0 for both branches (text disabled, DGTRS not loaded)

Text hierarchy (L1 / L2 / L3)

Each drone image has a VLM-generated caption (Qwen3-VL) split into 3 levels:

Level	Name	Content	Typical length
L1	Overview	First sentence of P1: land-cover summary with class percentages	15–30 tokens
L2	Full description	Complete P1 (inventory) + P2 (spatial layout with 5+ zones)	100–200 tokens
L3	Fingerprint	P3: unique landmarks and spatial signature for matching	20–50 tokens

All three levels are encoded by a single DGTRS-CLIP ViT-L-14 text encoder (248-token context via KPS positional embedding, 768-dim output) adapted with LoRA (rank=4 on Q/V in all 12 blocks).

Text fusion (shared MLP for both branches)

\mathbf{z}_{\text{text}} = \text{MLP}\bigl([\mathbf{z}_1 \;;\; \mathbf{z}_2 \;;\; \mathbf{z}_3]\bigr)

where [z₁ ; z₂ ; z₃] ∈ ℝ^{B×2304} is the concatenation of three 768-dim DGTRS-CLIP embeddings, and

\text{MLP}: \text{Linear}(2304, 1024) \to \text{GELU} \to \text{Linear}(1024, 1024), \quad \mathbf{z}_{\text{text}} \in \mathbb{R}^{B \times 1024}

Gated fusion (separate gates for query and gallery)

\mathbf{q} = \sigma(\alpha_q) \cdot \mathbf{d}_{\text{img}} + \bigl(1 - \sigma(\alpha_q)\bigr) \cdot \mathbf{d}_{\text{txt}} \qquad \text{(query branch)}

\mathbf{g} = \sigma(\alpha_g) \cdot \mathbf{s}_{\text{img}} + \bigl(1 - \sigma(\alpha_g)\bigr) \cdot \mathbf{s}_{\text{txt}} \qquad \text{(gallery branch)}

• α_q, α_g — separate learnable scalars in logit-space, init σ(α) ≈ 0.7
• σ — sigmoid function
• d_img, s_img ∈ ℝ^{B×1024} — DINOv3+MONA image embeddings
• d_txt, s_txt ∈ ℝ^{B×1024} — fused text embeddings
• For satellite images without captions: s_txt = None → g = s_img

Adaptation methods

Method	Applied to	Where	Params
MONA (CVPR 2025)	DINOv3 ViT-L/16 (drone + sat)	After MSA and MLP in each of 24 blocks	7.0M per encoder
LoRA (rank=4)	DGTRS-CLIP text encoder	Q and V projections in all 12 blocks	147K

MONA adapter (per block):

\mathbf{x} \leftarrow \mathbf{x} + \text{Up}_{64 \to 1024}\!\Bigl(\text{GELU}\bigl(\text{MonaOp}\bigl(\text{Down}_{1024 \to 64}(\hat{\mathbf{x}})\bigr)\bigr)\Bigr)

where x̂ = γ · LN(x) + γₓ · x (scaled LayerNorm, γ init 10⁻⁶, γₓ init 1)

\text{MonaOp}(\mathbf{x}) = \frac{\text{DWConv}_{3 \times 3}(\mathbf{x}) + \text{DWConv}_{5 \times 5}(\mathbf{x}) + \text{DWConv}_{7 \times 7}(\mathbf{x})}{3} + \mathbf{x}

LoRA (per attention layer):

\mathbf{Q}' = \mathbf{Q} + \frac{\alpha}{r} \cdot \mathbf{x} \mathbf{A}_Q^T \mathbf{B}_Q^T, \qquad \mathbf{V}' = \mathbf{V} + \frac{\alpha}{r} \cdot \mathbf{x} \mathbf{A}_V^T \mathbf{B}_V^T

where A ∈ ℝ^{r×d}, B ∈ ℝ^{d×r}, r = 4

Loss function

Symmetric InfoNCE with learnable temperature (CLIP-style logit_scale):

\mathcal{L} = w_{q \to g} \cdot \mathcal{L}_{q \to g} + w_{g \to q} \cdot \mathcal{L}_{g \to q}

\mathcal{L}_{q \to g} = \text{CrossEntropy}\!\left(\frac{\hat{\mathbf{q}} \cdot \hat{\mathbf{g}}^T}{\tau},\; \text{targets}\right), \qquad \mathcal{L}_{g \to q} = \text{CrossEntropy}\!\left(\frac{\hat{\mathbf{g}} \cdot \hat{\mathbf{q}}^T}{\tau},\; \text{targets}\right)

• τ = 1 / exp(logit_scale) — learnable scalar, clamped τ ∈ [0.01, 0.5], init τ₀ = 0.07
• w_{q→g} = 0.6, w_{g→q} = 0.4
• targets = [0, 1, 2, ..., B−1] — positives on diagonal
• label smoothing = 0.1

Loss and adapters run in fp32 (AMP autocast disabled) to prevent gradient overflow.

Metrics

Metric	Formula	Direction
R@K (Recall at K)	fraction of queries where correct gallery is in top-K	drone → satellite (primary)
Delta R@1	R@1(with_text) − R@1(baseline)	higher = text helps

Reported: R@1, R@5, R@10 for both q→g and g→q directions.

Optimizer & scheduler

Optimizer: AdamW
  - MONA adapters, TextFusionMLP, gate α_q, gate α_g, logit_scale:
      lr = 1e-4,  weight_decay = 1e-4
  - LoRA adapters (DGTRS-CLIP text encoder):
      lr = 1e-5  (10× lower, --text-lr-factor 0.1)

Scheduler: Linear warmup (2 epochs) + cosine annealing
  - Per-step (not per-epoch)
  - warmup: lr linearly ramps from 0 to lr_max over warmup_steps
  - cosine: lr decays from lr_max to 0 over remaining steps

Gradient clipping: max_norm = 1.0
Mixed precision: AMP fp16 for model forward, fp32 for loss

Augmentations

Transform	Drone (train)	Satellite (train)	Eval
RandomResizedCrop(256, scale=0.7–1.0)	✓	✓	—
Resize(256) + CenterCrop(256)	—	—	✓
RandomHorizontalFlip(0.5)	✓	✓	—
RandomRotation(15°)	✓	—	—
ColorJitter(0.3, 0.3, 0.2, 0.1)	✓	✓	—
RandomGrayscale(0.05)	✓	✓	—
GaussianBlur(k=3, σ=0.1–2.0)	✓	—	—
ImageNet Normalize	✓	✓	✓

Model summary

Component	Params	Trainable	Notes
DINOv3 ViT-L/16 LVD (drone)	303M	frozen	backbone weights frozen
MONA adapters (drone)	7.0M	7.0M	2 per block × 24 blocks, bottleneck=64
DINOv3 ViT-L/16 SAT (satellite)	303M	frozen	backbone weights frozen
MONA adapters (satellite)	7.0M	7.0M	2 per block × 24 blocks, bottleneck=64
DGTRS-CLIP ViT-L-14 (text)	124M	frozen	backbone weights frozen
LoRA adapters (text)	147K	147K	Q+V, rank=4, 12 blocks
TextFusionMLP (shared)	3.4M	3.4M	Linear(2304,1024) + GELU + Linear(1024,1024)
GatedFusion α_q + α_g	2	2	separate gate scalars
logit_scale	1	1	learnable temperature
Total	748M	17.6M (2.35%)	retrieval dim = 1024

Experiments

V3 — GTA-UAV + DINOv3 + DGTRS-CLIP (active)

Dataset: GTA-UAV-LR (33K drone + 14K satellite, GTA V synthetic)

• RGB: /home/servml/Документы/datasets/GTA-UAV-LR/
• Captions: /home/servml/Документы/datasets/GTA-UAV-LR-captions/ (40K JSON, 3-paragraph VLM)
• Segmentation: /home/servml/Документы/datasets/GTA-UAV-LR-aug/ (17 classes)
• Seg filter: 37,498 passed / 10,905 excluded (>=90% background+water)
• Split: 80/20 random (26,966 train / 6,742 test → 24,891/6,252 after seg filter)

V2 — UAV-GeoLoc + GeoRSCLIP (legacy)

Single backbone with template captions.

Query:   drone_img + caption -> GeoRSCLIP -> GatedFusion -> query
Gallery: sat_img -> GeoRSCLIP -> gallery

Dataset: UAV-GeoLoc Terrain split (206K train queries)

Structure

caption-test/
├── conf/                         # Gin configs (v2)
│   ├── balanced.gin
│   ├── baseline_no_text.gin
│   └── text_heavy.gin
├── nn_models/                    # Pre-trained checkpoints (v3, gitignored)
│   ├── DINO_WEB/                 # DINOv3 ViT-L/16 LVD-1689M (.pth)
│   ├── DINO_SAT/                 # DINOv3 ViT-L/16 SAT-493M (.safetensors)
│   └── LRSCLIP/                  # DGTRS-CLIP ViT-L-14 (.pt)
├── meta/                         # Generated metadata
│   ├── train_80.json             # 80% train split (26,966 pairs)
│   ├── test_20.json              # 20% test split (6,742 pairs)
│   └── seg_filter.json           # Segmentation filter results
├── scripts/
│   ├── make_split.py             # Create 80/20 train/test split
│   ├── filter_segmentation.py    # Exclude 90%+ background/water images
│   ├── compare_runs.py           # Delta R@1 comparison report
│   └── generate_captions.py      # Offline caption generation
├── src/
│   ├── datasets/
│   │   ├── gtauav_dataset.py     # GTA-UAV-LR loader + L1/L2/L3 parsing (v3)
│   │   └── visloc_with_captions.py  # UAV-GeoLoc loader (v2)
│   ├── models/
│   │   ├── dgtrs/                # Official DGTRS-CLIP text encoder (Apache-2.0)
│   │   │   ├── model.py          # DGTRSTextEncoder, build_model, tokenize
│   │   │   ├── simple_tokenizer.py  # BPE tokenizer (248 tokens)
│   │   │   └── bpe_simple_vocab_16e6.txt.gz
│   │   ├── asymmetric_encoder.py # DINOv3 + DGTRS + GatedFusion (v3)
│   │   └── dual_encoder.py       # GeoRSCLIP + GatedFusion (v2)
│   ├── losses/
│   │   └── multi_infonce.py      # InfoNCE with learnable temperature
│   ├── training/
│   │   ├── train_gtauav.py       # Training loop GTA-UAV (v3)
│   │   └── train.py              # Training loop UAV-GeoLoc (v2)
│   └── eval/
│       └── evaluate.py           # R@K metrics, Delta R@1
└── checkpoints/                  # GeoRSCLIP RS5M_ViT-B-32.pt (v2)

Prerequisites

torch>=2.0
safetensors
coloredlogs
tqdm
ftfy
regex
gin-config
Pillow
numpy

Workflow (V3 — GTA-UAV)

1. Create 80/20 split and filter segmentation

python -m scripts.make_split --output-dir meta
python -m scripts.filter_segmentation --output meta/seg_filter.json

2. Train baseline (no text)

python -m src.training.train_gtauav --baseline --filter-meta meta/seg_filter.json

3. Train with captions (L1/L2/L3)

python -m src.training.train_gtauav --filter-meta meta/seg_filter.json

4. Resume from checkpoint

python -m src.training.train_gtauav --resume out/gtauav/with_text/ckpt_epoch004.pt \
    --filter-meta meta/seg_filter.json

5. Compare and get verdict

python -m scripts.compare_runs \
    --baseline_report out/gtauav/baseline/eval_report.json \
    --full_report out/gtauav/with_text/eval_report.json \
    --output out/gtauav/comparison.md

Decision rule

Delta R@1 (drone→satellite)	Verdict
>= +3%	PASS — captions informative, proceed to NADEZHDA teacher
+1% to +3%	MARGINAL — add VLM refinement, re-run
0 to +1%	WEAK — redesign caption pipeline
< 0	HARMFUL — critical bug

Code style

• from __future__ import annotations everywhere
• Type hints on all signatures
• Google-style docstrings
• English-only comments