# Hard-negative Retrieval on KITTI object

Source: [Hard-negative Retrieval on KITTI object](<https://www.sota2.com/research/sota/hard-negative-retrieval-on-kitti-object>)

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

Data updated: 2026-03-31T10:31:07.695Z

Evaluation of Hard-negative retrieval performance on the KITTI object dataset, typically measured using Top-K accuracy and Margin scores.

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 | Hard-negative retrieval |
| Dataset (including any stated split) | KITTI object |
| Metric | Top-1 Acc |
| Unit | PERCENT |
| Direction | Higher is better |
| Results loaded | 6 of 6 |
| 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/hard-negative-retrieval-on-kitti-object.md?metric=Top-1%20Acc>): Higher is better; unit: PERCENT.

- [Top-5 Acc](<https://www.sota2.com/research/sota/hard-negative-retrieval-on-kitti-object.md?metric=Top-5%20Acc>): Higher is better; unit: PERCENT.

- [Hard-1 Acc](<https://www.sota2.com/research/sota/hard-negative-retrieval-on-kitti-object.md?metric=Hard-1%20Acc>): Higher is better; unit: PERCENT.

- [Hard-5 Acc](<https://www.sota2.com/research/sota/hard-negative-retrieval-on-kitti-object.md?metric=Hard-5%20Acc>): Higher is better; unit: PERCENT.

- [Margin Score](<https://www.sota2.com/research/sota/hard-negative-retrieval-on-kitti-object.md?metric=Margin%20Score>): Higher is better; unit: SCALAR.

## Recorded results

| Method | Top-1 Acc | Result date | Report paper |
| --- | --- | --- | --- |
| BINO | 65.2 | Not recorded | 69cb55340636e5bf4afca07c |
| CroCo-v2 | 56.3 | Not recorded | 69cb55340636e5bf4afca07c |
| CroCo-v1 | 45.1 | Not recorded | 69cb55340636e5bf4afca07c |
| DINOv3-style | 37.9 | Not recorded | 69cb55340636e5bf4afca07c |
| DINOv2-style | 14.5 | Not recorded | 69cb55340636e5bf4afca07c |
| iBOT-style | 12.4 | Not recorded | 69cb55340636e5bf4afca07c |

### BINO

Top-1 Acc: 65.2. Result ID: 69cba1d547f34ce924b48e49.

[Reporting paper](<https://www.sota2.com/research/paper/69cb55340636e5bf4afca07c.md>)

[Method origin paper](<https://www.sota2.com/research/paper/69cb55340636e5bf4afca07c.md>)

Recorded evaluation configuration: {"Evaluation Protocol":"Frozen no-linkage descriptor probe"}. This record does not assert independent reproduction.

### CroCo-v2

Top-1 Acc: 56.3. Result ID: 69cba1d547f34ce924b48e4b.

[Reporting paper](<https://www.sota2.com/research/paper/69cb55340636e5bf4afca07c.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa70541e209131cea4a451.md>)

[Implementation](<https://github.com/naver/croco>)

Recorded evaluation configuration: {"Evaluation Protocol":"Frozen no-linkage descriptor probe","Positional encoding":"RoPE"}. This record does not assert independent reproduction.

### CroCo-v1

Top-1 Acc: 45.1. Result ID: 69cba1d547f34ce924b48e4a.

[Reporting paper](<https://www.sota2.com/research/paper/69cb55340636e5bf4afca07c.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa6e591e209131cea3f3ea.md>)

[Implementation](<https://europe.naverlabs.com/research/computer-vision/croco/>)

Recorded evaluation configuration: {"Evaluation Protocol":"Frozen no-linkage descriptor probe","Positional encoding":"sincos"}. This record does not assert independent reproduction.

### DINOv3-style

Top-1 Acc: 37.9. Result ID: 69cba1d547f34ce924b48e4e.

[Reporting paper](<https://www.sota2.com/research/paper/69cb55340636e5bf4afca07c.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa7a931e209131cea83fc5.md>)

Recorded evaluation configuration: {"Evaluation Protocol":"Frozen no-linkage descriptor probe","Pre-training style":"DINOv3-style"}. This record does not assert independent reproduction.

### DINOv2-style

Top-1 Acc: 14.5. Result ID: 69cba1d547f34ce924b48e4d.

[Reporting paper](<https://www.sota2.com/research/paper/69cb55340636e5bf4afca07c.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa72181e209131cea546c3.md>)

[Implementation](<https://github.com/facebookresearch/dinov2>)

Recorded evaluation configuration: {"Evaluation Protocol":"Frozen no-linkage descriptor probe","Pre-training style":"DINOv2-style"}. This record does not assert independent reproduction.

### iBOT-style

Top-1 Acc: 12.4. Result ID: 69cba1d547f34ce924b48e4c.

[Reporting paper](<https://www.sota2.com/research/paper/69cb55340636e5bf4afca07c.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa6bb51e209131cea2fbe0.md>)

[Implementation](<https://github.com/bytedance/ibot>)

Recorded evaluation configuration: {"Evaluation Protocol":"Frozen no-linkage descriptor probe","Pre-training style":"iBOT-style"}. This record does not assert independent reproduction.

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