# SeCo-SBIR: Semantically Consistent Prompt Learning for Zero-Shot Sketch-Based Image Retrieval

arXiv: https://arxiv.org/abs/2608.03120 · HTML: https://mlsanity.com/paper/2608.03120/

## Key facts

- No citing papers in the map yet (data through 14 Oct 2026).
- Builds on 25 of its 49 references that are in the map.
- First posted on arXiv 4 Aug 2026.

## Builds on (most cited first, 25 of 25)

- [Deep Residual Learning for Image Recognition](https://mlsanity.com/paper/1512.03385/) (2015)
- [Learning Transferable Visual Models From Natural Language Supervision](https://mlsanity.com/paper/2103.00020/) (2021)
- [An Image is Worth 16x16 Words: Transformers for Image Recognition at Scale](https://mlsanity.com/paper/2010.11929/) (2020)
- [Representation Learning with Contrastive Predictive Coding](https://mlsanity.com/paper/1807.03748/) (2018)
- [FaceNet: A Unified Embedding for Face Recognition and Clustering](https://mlsanity.com/paper/1503.03832/) (2015)
- [Understanding Contrastive Representation Learning through Alignment and Uniformity on the Hypersphere](https://mlsanity.com/paper/2005.10242/) (2020)
- [DenseCLIP: Language-Guided Dense Prediction with Context-Aware Prompting](https://mlsanity.com/paper/2112.01518/) (2021)
- [Understanding the Behaviour of Contrastive Loss](https://mlsanity.com/paper/2012.09740/) (2020)
- [A Neural Representation of Sketch Drawings](https://mlsanity.com/paper/1704.03477/) (2017)
- [Visual-Language Prompt Tuning with Knowledge-guided Context Optimization](https://mlsanity.com/paper/2303.13283/) (2023)
- [Deep Metric Learning with Hierarchical Triplet Loss](https://mlsanity.com/paper/1810.06951/) (2018)
- [Doodle to Search: Practical Zero-Shot Sketch-based Image Retrieval](https://mlsanity.com/paper/1904.03451/) (2019)
- [CLIP for All Things Zero-Shot Sketch-Based Image Retrieval, Fine-Grained or Not](https://mlsanity.com/paper/2303.13440/) (2023)
- [A Zero-Shot Framework for Sketch-based Image Retrieval](https://mlsanity.com/paper/1807.11724/) (2018)
- [Deep Sketch Hashing: Fast Free-hand Sketch-Based Image Retrieval](https://mlsanity.com/paper/1703.05605/) (2017)
- [Zero-Shot Sketch-Image Hashing](https://mlsanity.com/paper/1803.02284/) (2018)
- [Do More Negative Samples Necessarily Hurt in Contrastive Learning?](https://mlsanity.com/paper/2205.01789/) (2022)
- [InfoNCE Loss Provably Learns Cluster-Preserving Representations](https://mlsanity.com/paper/2302.07920/) (2023)
- [StyLIP: Multi-Scale Style-Conditioned Prompt Learning for CLIP-based Domain Generalization](https://mlsanity.com/paper/2302.09251/) (2023)
- [Sketch3T: Test-Time Training for Zero-Shot SBIR](https://mlsanity.com/paper/2203.14691/) (2022)
- [A Theoretically Sound Upper Bound on the Triplet Loss for Improving the Efficiency of Deep Distance Metric Learning](https://mlsanity.com/paper/1904.08720/) (2019)
- [Zero-Shot Everything Sketch-Based Image Retrieval, and in Explainable Style](https://mlsanity.com/paper/2303.14348/) (2023)
- [Unknown Prompt, the only Lacuna: Unveiling CLIP's Potential for Open Domain Generalization](https://mlsanity.com/paper/2404.00710/) (2024)
- [Zero-Shot Sketch Based Image Retrieval using Graph Transformer](https://mlsanity.com/paper/2201.10185/) (2022)
- [SketchFusion: Learning Universal Sketch Features through Fusing Foundation Models](https://mlsanity.com/paper/2503.14129/) (2025)

## Abstract (arXiv)

Adapting CLIP for zero-shot sketch-based image retrieval (ZS-SBIR) via prompt learning faces a fundamental tension: the model must bridge the sketch-photo domain gap through task-specific adaptation, yet the added flexibility risks overfitting to seen training categories and eroding CLIP's zero-shot generalization. We present SeCo-SBIR, a semantically consistent prompt learning framework that resolves this tension from both sides. First, a text-guided multi-modal prompting strategy routes learnable prompt vectors through CLIP's text encoder and projects the resulting intermediate representations into the visual encoder at every layer via learnable coupling functions. Because the text encoder has already learned robust, abstract category-level semantics from large-scale language supervision, this mechanism injects transferable semantic knowledge directly into the visual pathway - adapting the model to the sketch-photo domain while inherently favoring generalization to unseen classes. Second, a perturbation-based consistency constraint addresses the residual overfitting risk from the learnable coupling functions by aligning the adapted model with a frozen CLIP reference branch using an asymmetric InfoNCE objective - augmented inputs feed the frozen branch while clean inputs feed the trainable branch - anchoring the learned representations to CLIP's generalizable feature space. Together with lightweight adapters and a multi-objective loss combining triplet, NT-Xent, and classification terms, SeCo-SBIR achieves state-of-the-art results on all three standard ZS-SBIR benchmarks across categorical, generalized, and across-dataset settings.