# Cross-category Fine-Grained Zero-Shot Sketch-Based Image Retrieval on Sketchy

Source: [Cross-category Fine-Grained Zero-Shot Sketch-Based Image Retrieval on Sketchy](<https://www.sota2.com/research/sota/cross-category-fine-grained-zero-shot-sketch-based-image-retrieval-on-sketchy>)

Collection index: [llms.txt](<https://www.sota2.com/research/llms.txt>)

Data updated: 2026-02-26T06:06:40.913Z

Evaluation of Cross-category Fine-Grained Zero-Shot Sketch-Based Image Retrieval performance on the Sketchy dataset, measured using Top-1 and Top-5 accuracy.

A research result applies to its stated dataset, split, metric and evaluation configuration. Compare like-for-like settings, preserve the metric's unit and direction, and distinguish paper-reported results from independent reproduction. The best result recorded here is not a claim of exhaustive, current global coverage.

| Scope | Recorded value |
| --- | --- |
| Task | Cross-category Fine-Grained Zero-Shot Sketch-Based Image Retrieval |
| Dataset (including any stated split) | Sketchy |
| Metric | Top-1 Acc |
| Unit | PERCENT |
| Direction | Higher is better |
| Results loaded | 9 of 9 |
| Coverage | All results currently returned for this benchmark |

Evaluation settings may differ between results. The original papers and recorded configurations are needed to establish comparability. A shared metric name alone does not establish an identical protocol.

## Metrics

- [Top-1 Acc](<https://www.sota2.com/research/sota/cross-category-fine-grained-zero-shot-sketch-based-image-retrieval-on-sketchy.md?metric=Top-1%20Acc>): Higher is better; unit: PERCENT.

- [Top-5 Acc](<https://www.sota2.com/research/sota/cross-category-fine-grained-zero-shot-sketch-based-image-retrieval-on-sketchy.md?metric=Top-5%20Acc>): Higher is better; unit: PERCENT.

## Recorded results

| Method | Top-1 Acc | Result date | Report paper |
| --- | --- | --- | --- |
| CLIP for FG-ZS-SBIR (Ours) | 28.68 | Not recorded | 68fa6edf1e209131cea428d4 |
| B-FG-Deep | 27.62 | Not recorded | 68fa6edf1e209131cea428d4 |
| B-FG-MM | 27.16 | Not recorded | 68fa6edf1e209131cea428d4 |
| B-FG-IP | 26.69 | Not recorded | 68fa6edf1e209131cea428d4 |
| B-FG-Cond | 25.98 | Not recorded | 68fa6edf1e209131cea428d4 |
| CC-DG | 22.6 | Not recorded | 68fa6edf1e209131cea428d4 |
| B-FG-Lin | 15.75 | Not recorded | 68fa6edf1e209131cea428d4 |
| Cross-GRAD | 13.4 | Not recorded | 68fa6edf1e209131cea428d4 |
| B-FG-FT | 1.23 | Not recorded | 68fa6edf1e209131cea428d4 |

### CLIP for FG-ZS-SBIR (Ours)

Top-1 Acc: 28.68. Result ID: 699076da3a6c31484900c101.

[Reporting paper](<https://www.sota2.com/research/paper/68fa6edf1e209131cea428d4.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa6edf1e209131cea428d4.md>)

[Implementation](<https://aneeshan95.github.io/Sketch_LVM/>)

Recorded evaluation configuration: {"Backbone":"ViT-B/16 (CLIP)","Prompting Strategy":"Fine-tuned Layer-Norm prompts"}. This record does not assert independent reproduction.

### B-FG-Deep

Top-1 Acc: 27.62. Result ID: 699076da3a6c31484900c100.

[Reporting paper](<https://www.sota2.com/research/paper/68fa6edf1e209131cea428d4.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa6d181e209131cea37ae9.md>)

[Implementation](<https://www.sota2.com/github.com/kmnp/vpt>)

Recorded evaluation configuration: {"Backbone":"ViT-B/16 (CLIP)","Prompting Strategy":"Deep Prompting"}. This record does not assert independent reproduction.

### B-FG-MM

Top-1 Acc: 27.16. Result ID: 699076da3a6c31484900c0ff.

[Reporting paper](<https://www.sota2.com/research/paper/68fa6edf1e209131cea428d4.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa6ef21e209131cea4311c.md>)

[Implementation](<https://github.com/muzairkhattak/multimodal-prompt-learning>)

Recorded evaluation configuration: {"Backbone":"ViT-B/16 (CLIP)","Prompting Strategy":"Multi-modal Prompting"}. This record does not assert independent reproduction.

### B-FG-IP

Top-1 Acc: 26.69. Result ID: 699076da3a6c31484900c0fe.

[Reporting paper](<https://www.sota2.com/research/paper/68fa6edf1e209131cea428d4.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa6d8b1e209131cea3a814.md>)

[Implementation](<https://github.com/KaiyangZhou/CoOp>)

Recorded evaluation configuration: {"Backbone":"ViT-B/16 (CLIP)","Prompting Strategy":"Independent Prompting"}. This record does not assert independent reproduction.

### B-FG-Cond

Top-1 Acc: 25.98. Result ID: 699076da3a6c31484900c0fd.

[Reporting paper](<https://www.sota2.com/research/paper/68fa6edf1e209131cea428d4.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa6d8c1e209131cea3a831.md>)

[Implementation](<https://github.com/KaiyangZhou/CoOp>)

Recorded evaluation configuration: {"Backbone":"ViT-B/16 (CLIP)","Prompting Strategy":"Conditional Prompting (ResNet-18)"}. This record does not assert independent reproduction.

### CC-DG

Top-1 Acc: 22.6. Result ID: 699076da3a6c31484900c0fa.

[Reporting paper](<https://www.sota2.com/research/paper/68fa6edf1e209131cea428d4.md>)

Recorded evaluation configuration: {"Evaluation Protocol":"Cross-category FG-ZS-SBIR"}. This record does not assert independent reproduction.

### B-FG-Lin

Top-1 Acc: 15.75. Result ID: 699076da3a6c31484900c0fc.

[Reporting paper](<https://www.sota2.com/research/paper/68fa6edf1e209131cea428d4.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa69581e209131cea21ad0.md>)

[Implementation](<https://github.com/OpenAI/CLIP>)

Recorded evaluation configuration: {"Backbone":"ViT-B/16 (CLIP)","Prompting Strategy":"Linear Probe"}. This record does not assert independent reproduction.

### Cross-GRAD

Top-1 Acc: 13.4. Result ID: 699076da3a6c31484900c0f9.

[Reporting paper](<https://www.sota2.com/research/paper/68fa6edf1e209131cea428d4.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa644f1e209131cea03efd.md>)

[Implementation](<https://github.com/vihari/crossgrad>)

Recorded evaluation configuration: {"Evaluation Protocol":"Cross-category FG-ZS-SBIR"}. This record does not assert independent reproduction.

### B-FG-FT

Top-1 Acc: 1.23. Result ID: 699076da3a6c31484900c0fb.

[Reporting paper](<https://www.sota2.com/research/paper/68fa6edf1e209131cea428d4.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa69581e209131cea21ad0.md>)

[Implementation](<https://github.com/OpenAI/CLIP>)

Recorded evaluation configuration: {"Backbone":"ViT-B/16 (CLIP)","Prompting Strategy":"Fine-tune Image Encoder"}. This record does not assert independent reproduction.

Page 1 of 1. This is an index page, not the full collection.
