# Line Segment Detection on ShanghaiTech v1.0 (test)

Source: [Line Segment Detection on ShanghaiTech v1.0 (test)](<https://www.sota2.com/research/sota/line-segment-detection-on-shanghaitech-v1-0-test>)

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

Data updated: 2026-02-26T06:11:18.729Z

Evaluation of Line Segment Detection performance on the test split of the ShanghaiTech v1.0 dataset, reporting metrics such as sAP5, sAP10, APH, and inference speed (FPS).

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 | Line Segment Detection |
| Dataset (including any stated split) | ShanghaiTech v1.0 (test) |
| Metric | sAP5 |
| Unit | PERCENT |
| Direction | Higher is better |
| Results loaded | 15 of 15 |
| 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

- [sAP5](<https://www.sota2.com/research/sota/line-segment-detection-on-shanghaitech-v1-0-test.md?metric=sAP5>): Higher is better; unit: PERCENT.

- [sAP10](<https://www.sota2.com/research/sota/line-segment-detection-on-shanghaitech-v1-0-test.md?metric=sAP10>): Higher is better; unit: PERCENT.

- [APH](<https://www.sota2.com/research/sota/line-segment-detection-on-shanghaitech-v1-0-test.md?metric=APH>): Higher is better; unit: PERCENT.

- [FPS (batch-size=1)](<https://www.sota2.com/research/sota/line-segment-detection-on-shanghaitech-v1-0-test.md?metric=FPS%20(batch-size%3D1)>): Higher is better; unit: SCALAR.

- [FPS (batch-size=10)](<https://www.sota2.com/research/sota/line-segment-detection-on-shanghaitech-v1-0-test.md?metric=FPS%20(batch-size%3D10)>): Higher is better; unit: SCALAR.

## Recorded results

| Method | sAP5 | Result date | Report paper |
| --- | --- | --- | --- |
| F-Clip (HR) | 64.3 | Not recorded | 68fa6e4a1e209131cea3f04f |
| HAWP (re-trained) | 63.1 | Not recorded | 68fa6e4a1e209131cea3f04f |
| F-Clip (HG2-LB) | 62.6 | Not recorded | 68fa6e4a1e209131cea3f04f |
| HAWP | 62.5 | Not recorded | 68fa6e4a1e209131cea3f04f |
| TP-LSD (re-trained) | 62.4 | Not recorded | 68fa6e4a1e209131cea3f04f |
| F-Clip (HG2) | 61.3 | Not recorded | 68fa6e4a1e209131cea3f04f |
| L-CNN | 58.9 | Not recorded | 68fa6e4a1e209131cea3f04f |
| F-Clip (HG1) | 58.6 | Not recorded | 68fa6e4a1e209131cea3f04f |
| F-Clip (HG1-D3) | 57.8 | Not recorded | 68fa6e4a1e209131cea3f04f |
| TP-LSD (Res34) | 57.5 | Not recorded | 68fa6e4a1e209131cea3f04f |
| TP-LSD (Res34-Lite) | 56.4 | Not recorded | 68fa6e4a1e209131cea3f04f |
| F-Clip (HG1-D2) | 52.7 | Not recorded | 68fa6e4a1e209131cea3f04f |
| AFM | 18.5 | Not recorded | 68fa6e4a1e209131cea3f04f |
| LSD (320) | 6.7 | Not recorded | 68fa6e4a1e209131cea3f04f |
| DWP | 3.7 | Not recorded | 68fa6e4a1e209131cea3f04f |

### F-Clip (HR)

sAP5: 64.3. Result ID: 699028f33a6c314849fe8a8a.

[Reporting paper](<https://www.sota2.com/research/paper/68fa6e4a1e209131cea3f04f.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa6e4a1e209131cea3f04f.md>)

[Implementation](<https://github.com/Delay-Xili/F-Clip>)

Recorded evaluation configuration: {"Category":"One-stage","Backbone":"HR"}. This record does not assert independent reproduction.

### HAWP (re-trained)

sAP5: 63.1. Result ID: 699028f33a6c314849fe8a81.

[Reporting paper](<https://www.sota2.com/research/paper/68fa6e4a1e209131cea3f04f.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa67241e209131cea14a50.md>)

Recorded evaluation configuration: {"Category":"Two-stage","Retrained":true}. This record does not assert independent reproduction.

### F-Clip (HG2-LB)

sAP5: 62.6. Result ID: 699028f33a6c314849fe8a89.

[Reporting paper](<https://www.sota2.com/research/paper/68fa6e4a1e209131cea3f04f.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa6e4a1e209131cea3f04f.md>)

[Implementation](<https://github.com/Delay-Xili/F-Clip>)

Recorded evaluation configuration: {"Category":"One-stage","Backbone":"HG2","Variant":"LB"}. This record does not assert independent reproduction.

### HAWP

sAP5: 62.5. Result ID: 699028f33a6c314849fe8a80.

[Reporting paper](<https://www.sota2.com/research/paper/68fa6e4a1e209131cea3f04f.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa67241e209131cea14a50.md>)

Recorded evaluation configuration: {"Category":"Two-stage"}. This record does not assert independent reproduction.

### TP-LSD (re-trained)

sAP5: 62.4. Result ID: 699028f33a6c314849fe8a84.

[Reporting paper](<https://www.sota2.com/research/paper/68fa6e4a1e209131cea3f04f.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa68511e209131cea1b6df.md>)

Recorded evaluation configuration: {"Category":"One-stage","Retrained":true}. This record does not assert independent reproduction.

### F-Clip (HG2)

sAP5: 61.3. Result ID: 699028f33a6c314849fe8a88.

[Reporting paper](<https://www.sota2.com/research/paper/68fa6e4a1e209131cea3f04f.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa6e4a1e209131cea3f04f.md>)

[Implementation](<https://github.com/Delay-Xili/F-Clip>)

Recorded evaluation configuration: {"Category":"One-stage","Backbone":"HG2"}. This record does not assert independent reproduction.

### L-CNN

sAP5: 58.9. Result ID: 699028f33a6c314849fe8a7f.

[Reporting paper](<https://www.sota2.com/research/paper/68fa6e4a1e209131cea3f04f.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa69ea1e209131cea24de0.md>)

[Implementation](<https://github.com/zhou13/lcnn>)

Recorded evaluation configuration: {"Category":"Two-stage"}. This record does not assert independent reproduction.

### F-Clip (HG1)

sAP5: 58.6. Result ID: 699028f33a6c314849fe8a87.

[Reporting paper](<https://www.sota2.com/research/paper/68fa6e4a1e209131cea3f04f.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa6e4a1e209131cea3f04f.md>)

[Implementation](<https://github.com/Delay-Xili/F-Clip>)

Recorded evaluation configuration: {"Category":"One-stage","Backbone":"HG1"}. This record does not assert independent reproduction.

### F-Clip (HG1-D3)

sAP5: 57.8. Result ID: 699028f33a6c314849fe8a86.

[Reporting paper](<https://www.sota2.com/research/paper/68fa6e4a1e209131cea3f04f.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa6e4a1e209131cea3f04f.md>)

[Implementation](<https://github.com/Delay-Xili/F-Clip>)

Recorded evaluation configuration: {"Category":"One-stage","Backbone":"HG1","Variant":"D3"}. This record does not assert independent reproduction.

### TP-LSD (Res34)

sAP5: 57.5. Result ID: 699028f33a6c314849fe8a83.

[Reporting paper](<https://www.sota2.com/research/paper/68fa6e4a1e209131cea3f04f.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa68511e209131cea1b6df.md>)

Recorded evaluation configuration: {"Category":"One-stage","Backbone":"ResNet-34"}. This record does not assert independent reproduction.

### TP-LSD (Res34-Lite)

sAP5: 56.4. Result ID: 699028f33a6c314849fe8a82.

[Reporting paper](<https://www.sota2.com/research/paper/68fa6e4a1e209131cea3f04f.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa68511e209131cea1b6df.md>)

Recorded evaluation configuration: {"Category":"One-stage","Backbone":"ResNet-34","Variant":"Lite"}. This record does not assert independent reproduction.

### F-Clip (HG1-D2)

sAP5: 52.7. Result ID: 699028f33a6c314849fe8a85.

[Reporting paper](<https://www.sota2.com/research/paper/68fa6e4a1e209131cea3f04f.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa6e4a1e209131cea3f04f.md>)

[Implementation](<https://github.com/Delay-Xili/F-Clip>)

Recorded evaluation configuration: {"Category":"One-stage","Backbone":"HG1","Variant":"D2"}. This record does not assert independent reproduction.

### AFM

sAP5: 18.5. Result ID: 699028f33a6c314849fe8a7d.

[Reporting paper](<https://www.sota2.com/research/paper/68fa6e4a1e209131cea3f04f.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa65721e209131cea0a7e8.md>)

Recorded evaluation configuration: {"Category":"Two-stage"}. This record does not assert independent reproduction.

### LSD (320)

sAP5: 6.7. Result ID: 699028f33a6c314849fe8a7c.

[Reporting paper](<https://www.sota2.com/research/paper/68fa6e4a1e209131cea3f04f.md>)

Recorded evaluation configuration: {"Category":"Two-stage","Input Resolution":"320"}. This record does not assert independent reproduction.

### DWP

sAP5: 3.7. Result ID: 699028f33a6c314849fe8a7e.

[Reporting paper](<https://www.sota2.com/research/paper/68fa6e4a1e209131cea3f04f.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa67f41e209131cea19652.md>)

Recorded evaluation configuration: {"Category":"Two-stage"}. This record does not assert independent reproduction.

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