# When Does Geometric View Synthesis Help Wine Label Retrieval? A Public One-Shot Benchmark Across Self-Supervised and Vision-Language Backbones

arXiv: https://arxiv.org/abs/2609.33359 · HTML: https://mlsanity.com/paper/2609.33359/

## Key facts

- No citing papers in the map yet (data through 14 Oct 2026).
- Builds on 17 of its 27 references that are in the map.
- First posted on arXiv 27 Sep 2026.

## Builds on (most cited first, 17 of 17)

- [Segment Anything](https://mlsanity.com/paper/2304.02643/) (2023)
- [DINOv2: Learning Robust Visual Features without Supervision](https://mlsanity.com/paper/2304.07193/) (2023)
- [Emerging Properties in Self-Supervised Vision Transformers](https://mlsanity.com/paper/2104.14294/) (2021)
- [Grounding DINO: Marrying DINO with Grounded Pre-Training for Open-Set Object Detection](https://mlsanity.com/paper/2303.05499/) (2023)
- [Billion-scale similarity search with GPUs](https://mlsanity.com/paper/1702.08734/) (2017)
- [SuperPoint: Self-Supervised Interest Point Detection and Description](https://mlsanity.com/paper/1712.07629/) (2017)
- [SigLIP 2: Multilingual Vision-Language Encoders with Improved Semantic Understanding, Localization, and Dense Features](https://mlsanity.com/paper/2502.14786/) (2025)
- [LoFTR: Detector-Free Local Feature Matching with Transformers](https://mlsanity.com/paper/2104.00680/) (2021)
- [LightGlue: Local Feature Matching at Light Speed](https://mlsanity.com/paper/2306.13643/) (2023)
- [ILIAS: Instance-Level Image retrieval At Scale](https://mlsanity.com/paper/2502.11748/) (2025)
- [AMES: Asymmetric and Memory-Efficient Similarity Estimation for Instance-level Retrieval](https://mlsanity.com/paper/2408.03282/) (2024)
- [Multi-Modal Reasoning Graph for Scene-Text Based Fine-Grained Image Classification and Retrieval](https://mlsanity.com/paper/2009.09809/) (2020)
- [Integrating Scene Text and Visual Appearance for Fine-Grained Image Classification](https://mlsanity.com/paper/1704.04613/) (2017)
- [Fine-grained Image Classification and Retrieval by Combining Visual and Locally Pooled Textual Features](https://mlsanity.com/paper/2001.04732/) (2020)
- [Learning to Taste: A Multimodal Wine Dataset](https://mlsanity.com/paper/2308.16900/) (2023)
- [Domain invariant hierarchical embedding for grocery products recognition](https://mlsanity.com/paper/1902.00760/) (2019)
- [Bridging the Catalog-to-Real Gap: Scalable Product Recognition via Multi-Stage Contrastive Learning](https://mlsanity.com/paper/2607.09888/) (2026)

## Abstract (arXiv)

Geometric view synthesis can expand a single wine-label photograph into a training set, but its value with pretrained image encoders is unclear. We study this on a public WineSensed-derived benchmark of 1,000 classes, one enrollment photograph per class, and 4,295 real queries. With the earlier DINO vision transformer (ViT-S/16) recipe, geometric views raise top-1 accuracy from 34.1% to 62.6-63.7%, about three times the gain from two-dimensional (2D) augmentation. Frozen SigLIP 2-B already reaches 94.7%. A linear head over its frozen features gains 1.2-1.3 percentage points with the two geometric pipelines localized by the Segment Anything Model (SAM), while the other pipelines gain an inconclusive 0.3-0.6 points. Low-rank adaptation (LoRA) and validation-selected full fine-tuning show no clear gain within the reported confidence intervals; fixed-budget full fine-tuning loses 9-24 points. SAM localization supplies all six views for 99% of sources, compared with 43% for the edge-based front end. Recognition differences between the two cylinder constructions depend on the training recipe and are confounded by their crop and canvas conventions. Rendered-cylinder tests show different responses to source tilt, but an uncalibrated rim-ratio proxy establishes no corresponding trend in recognition on real photographs. An author-confirmed audit of 50 residual errors identifies 21 query-enrollment appearance mismatches, without establishing an irreducible error rate. These results support geometric synthesis for the tested self-supervised recipe and a smaller benefit through frozen-feature adaptation of the text-supervised encoder.