# Multimodal Violence Detection on NTU-CCTV (test)

Source: [Multimodal Violence Detection on NTU-CCTV (test)](<https://www.sota2.com/research/sota/multimodal-violence-detection-on-ntu-cctv-test>)

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

Data updated: 2026-04-07T07:45:54.006Z

Evaluation of Multimodal Violence Detection performance on the test split of the NTU-CCTV dataset, using Accuracy and various F1 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 | Multimodal Violence Detection |
| Dataset (including any stated split) | NTU-CCTV (test) |
| Metric | Accuracy |
| Unit | PERCENT |
| Direction | Higher is better |
| Results loaded | 17 of 17 |
| 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

- [Accuracy](<https://www.sota2.com/research/sota/multimodal-violence-detection-on-ntu-cctv-test.md?metric=Accuracy>): Higher is better; unit: PERCENT.

- [F1 (Violence)](<https://www.sota2.com/research/sota/multimodal-violence-detection-on-ntu-cctv-test.md?metric=F1%20(Violence)>): Higher is better; unit: PERCENT.

- [F1 (Non-Violence)](<https://www.sota2.com/research/sota/multimodal-violence-detection-on-ntu-cctv-test.md?metric=F1%20(Non-Violence)>): Higher is better; unit: PERCENT.

- [Macro F1](<https://www.sota2.com/research/sota/multimodal-violence-detection-on-ntu-cctv-test.md?metric=Macro%20F1>): Higher is better; unit: PERCENT.

## Recorded results

| Method | Accuracy | Result date | Report paper |
| --- | --- | --- | --- |
| CoLoRSMamba | 88.63 | Not recorded | 69d48f997d0172140ec74358 |
| LAVISH | 86.44 | Not recorded | 69d48f997d0172140ec74358 |
| MBT | 85.26 | Not recorded | 69d48f997d0172140ec74358 |
| CAV-MAE Sync | 84.92 | Not recorded | 69d48f997d0172140ec74358 |
| DBVideoMamba | 83.99 | Not recorded | 69d48f997d0172140ec74358 |
| TIM | 83.07 | Not recorded | 69d48f997d0172140ec74358 |
| VideoMamba-M | 82.22 | Not recorded | 69d48f997d0172140ec74358 |
| CUE-Net | 82.06 | Not recorded | 69d48f997d0172140ec74358 |
| VideoMambaPro-M | 80.96 | Not recorded | 69d48f997d0172140ec74358 |
| UniFormer-V2 | 80.37 | Not recorded | 69d48f997d0172140ec74358 |
| SlowFast-AV | 79.11 | Not recorded | 69d48f997d0172140ec74358 |
| VideoSwin-B | 78.77 | Not recorded | 69d48f997d0172140ec74358 |
| DepMamba | 78.18 | Not recorded | 69d48f997d0172140ec74358 |
| SlowFast | 76.41 | Not recorded | 69d48f997d0172140ec74358 |
| AudioMamba | 72.79 | Not recorded | 69d48f997d0172140ec74358 |
| AST | 71.02 | Not recorded | 69d48f997d0172140ec74358 |
| Resnet-50 | 69.17 | Not recorded | 69d48f997d0172140ec74358 |

### CoLoRSMamba

Accuracy: 88.63. Result ID: 69d4b5e547f34ce924bc3c19.

[Reporting paper](<https://www.sota2.com/research/paper/69d48f997d0172140ec74358.md>)

[Method origin paper](<https://www.sota2.com/research/paper/69d48f997d0172140ec74358.md>)

Recorded evaluation configuration: {"Architecture":"SSM","Modality":"audio-video"}. This record does not assert independent reproduction.

### LAVISH

Accuracy: 86.44. Result ID: 69d4b5e547f34ce924bc3c16.

[Reporting paper](<https://www.sota2.com/research/paper/69d48f997d0172140ec74358.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa6f081e209131cea436f3.md>)

[Implementation](<https://genjib.github.io/project_page/LAVISH/>)

Recorded evaluation configuration: {"Architecture":"Transformer","Modality":"audio-video"}. This record does not assert independent reproduction.

### MBT

Accuracy: 85.26. Result ID: 69d4b5e547f34ce924bc3c15.

[Reporting paper](<https://www.sota2.com/research/paper/69d48f997d0172140ec74358.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa6e221e209131cea3e0de.md>)

Recorded evaluation configuration: {"Architecture":"Transformer","Modality":"audio-video"}. This record does not assert independent reproduction.

### CAV-MAE Sync

Accuracy: 84.92. Result ID: 69d4b5e547f34ce924bc3c17.

[Reporting paper](<https://www.sota2.com/research/paper/69d48f997d0172140ec74358.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa78b41e209131cea79dbd.md>)

Recorded evaluation configuration: {"Architecture":"Transformer","Modality":"audio-video"}. This record does not assert independent reproduction.

### DBVideoMamba

Accuracy: 83.99. Result ID: 69d4b5e547f34ce924bc3c12.

[Reporting paper](<https://www.sota2.com/research/paper/69d48f997d0172140ec74358.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa7b901e209131cea89264.md>)

Recorded evaluation configuration: {"Architecture":"SSM","Modality":"video-only"}. This record does not assert independent reproduction.

### TIM

Accuracy: 83.07. Result ID: 69d4b5e547f34ce924bc3c14.

[Reporting paper](<https://www.sota2.com/research/paper/69d48f997d0172140ec74358.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa72cf1e209131cea58a19.md>)

[Implementation](<https://github.com/JacobChalk/TIM>)

Recorded evaluation configuration: {"Architecture":"Transformer","Modality":"audio-video"}. This record does not assert independent reproduction.

### VideoMamba-M

Accuracy: 82.22. Result ID: 69d4b5e547f34ce924bc3c10.

[Reporting paper](<https://www.sota2.com/research/paper/69d48f997d0172140ec74358.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa72831e209131cea56ea1.md>)

[Implementation](<https://github.com/OpenGVLab/VideoMamba>)

Recorded evaluation configuration: {"Architecture":"SSM","Modality":"video-only"}. This record does not assert independent reproduction.

### CUE-Net

Accuracy: 82.06. Result ID: 69d4b5e547f34ce924bc3c0f.

[Reporting paper](<https://www.sota2.com/research/paper/69d48f997d0172140ec74358.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa73091e209131cea59fa2.md>)

Recorded evaluation configuration: {"Architecture":"CNN+Transformer","Modality":"video-only"}. This record does not assert independent reproduction.

### VideoMambaPro-M

Accuracy: 80.96. Result ID: 69d4b5e547f34ce924bc3c11.

[Reporting paper](<https://www.sota2.com/research/paper/69d48f997d0172140ec74358.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa75bb1e209131cea6998c.md>)

Recorded evaluation configuration: {"Architecture":"SSM","Modality":"video-only"}. This record does not assert independent reproduction.

### UniFormer-V2

Accuracy: 80.37. Result ID: 69d4b5e547f34ce924bc3c0e.

[Reporting paper](<https://www.sota2.com/research/paper/69d48f997d0172140ec74358.md>)

Recorded evaluation configuration: {"Architecture":"CNN+Transformer","Modality":"video-only"}. This record does not assert independent reproduction.

### SlowFast-AV

Accuracy: 79.11. Result ID: 69d4b5e547f34ce924bc3c13.

[Reporting paper](<https://www.sota2.com/research/paper/69d48f997d0172140ec74358.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa673a1e209131cea15085.md>)

Recorded evaluation configuration: {"Architecture":"CNN","Modality":"audio-video"}. This record does not assert independent reproduction.

### VideoSwin-B

Accuracy: 78.77. Result ID: 69d4b5e547f34ce924bc3c0d.

[Reporting paper](<https://www.sota2.com/research/paper/69d48f997d0172140ec74358.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa6a391e209131cea26be4.md>)

[Implementation](<https://github.com/SwinTransformer/Video-Swin-Transformer>)

Recorded evaluation configuration: {"Architecture":"Transformer","Modality":"video-only"}. This record does not assert independent reproduction.

### DepMamba

Accuracy: 78.18. Result ID: 69d4b5e547f34ce924bc3c18.

[Reporting paper](<https://www.sota2.com/research/paper/69d48f997d0172140ec74358.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa74fc1e209131cea6515f.md>)

Recorded evaluation configuration: {"Architecture":"SSM","Modality":"audio-video"}. This record does not assert independent reproduction.

### SlowFast

Accuracy: 76.41. Result ID: 69d4b5e547f34ce924bc3c0c.

[Reporting paper](<https://www.sota2.com/research/paper/69d48f997d0172140ec74358.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa673a1e209131cea15085.md>)

Recorded evaluation configuration: {"Architecture":"CNN","Modality":"video-only"}. This record does not assert independent reproduction.

### AudioMamba

Accuracy: 72.79. Result ID: 69d4b5e547f34ce924bc3c0b.

[Reporting paper](<https://www.sota2.com/research/paper/69d48f997d0172140ec74358.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa73841e209131cea5ce25.md>)

Recorded evaluation configuration: {"Architecture":"SSM","Modality":"audio-only"}. This record does not assert independent reproduction.

### AST

Accuracy: 71.02. Result ID: 69d4b5e547f34ce924bc3c0a.

[Reporting paper](<https://www.sota2.com/research/paper/69d48f997d0172140ec74358.md>)

[Method origin paper](<https://www.sota2.com/research/paper/68fa6a521e209131cea27512.md>)

[Implementation](<https://github.com/YuanGongND/ast>)

Recorded evaluation configuration: {"Architecture":"Transformer","Modality":"audio-only"}. This record does not assert independent reproduction.

### Resnet-50

Accuracy: 69.17. Result ID: 69d4b5e547f34ce924bc3c09.

[Reporting paper](<https://www.sota2.com/research/paper/69d48f997d0172140ec74358.md>)

Recorded evaluation configuration: {"Architecture":"CNN","Modality":"audio-only"}. This record does not assert independent reproduction.

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