From Weapons to Events: Evaluating Evidential Support in Multimodal AI for Policing
Organizations: Centre for Vision, Speech and Signal Processing, University of Surrey, Guildford, United Kingdom · AI Fellow, Incubator for AI (i.AI), Department for Digital, Culture, Media and Sport, London, United Kingdom
Abstract
Computer vision (CV) for policing has centred on recognising objects such as seized weapons, while body-worn video (BWV) is still reviewed by people, with existing AI tools using only its audio. Multimodal video AI could make BWV review faster and cheaper. At the object level, we present the KnifeHunter dataset, developed with UK law enforcement, with over 25,000 images of 543 knife classes and a protocol reflecting investigative conditions, and KHNet, a retrieval architecture fusing local and global features, which with about 125,000 distractor images attains 85.7% and 63.4% mean Average Precision (mAP) on the medium and hard protocols. At the event level, we present an evaluation protocol for video understanding pipelines, which describe a video as events and a transcript, that scores transcription, speaker attribution, event detection, temporal localisation and unsupported events. Whether each event is supported by the recording is labelled by a judge validated against blinded human labels. We apply it to complete pipelines, three 2026 omni-modal models and our hybrid NarrativeBridge-PAL-GPT-5 pipeline, on UKBWV-27, a pilot stress-test set of 27 publicly released UK BWV recordings (118 minutes). On this set the hybrid pipeline leads on five of six measures, yet it leaves 13.6% of predicted events unsupported and misattributes 27.7% of utterances, and every pipeline degrades on recordings from edited releases, long recordings and noisy audio. We propose readiness levels separating assistive tasks from those requiring verification of every output.
Figures & tables
| Aggregation | Medium | Hard | Medium+D | Hard+D |
|---|---|---|---|---|
| GeMNet ( Radenović et al., 2019 ) | 83.1 | 61.2 | 81.5 | 57.9 |
| ActNet ( Husain et al., 2021 ) | 83.9 | 62.1 | 82.6 | 58.9 |
| TokenNet ( Wu et al., 2022 ) | 84.8 | 62.5 | 83.5 | 59.8 |
| SENet ( Lee et al., 2023 ) | 85.6 | 63.2 | 84.3 | 61.5 |
| KHNet (ours) | 86.5 | 65.3 | 85.7 | 63.4 |
| Model | WER % | Speaker % | F1 | mIoU | Unsup. % | Latency |
|---|---|---|---|---|---|---|
| Hybrid pipeline (ours) | 5.7 | 72.3 | 63.1 | 0.44 | 13.6 | 37.3 |
| Gemini 3.1 Pro | 7.9 | 69.2 | 60.6 | 0.35 | 14.2 | 38.5 |
| Qwen3.5-Omni-Plus | 12.5 | 63.8 | 52.0 | 0.41 | 21.7 | 24.1 |
| Nemotron 3 Nano Omni | 13.7 | 52.4 | 49.7 | 0.38 | 29.5 | 29.8 |
| NarrativeBridge + GPT-5 (C1) | n/a | n/a | 56.5 | 0.39 | 13.9 | 37.9 |
| Condition | Videos | Hybrid | Gemini | Qwen | Nemotron |
| Released as pure BWV | 22 | 64.7 | 62.4 | 54.0 | 51.6 |
| Released with overlays | 3 | 55.4 | 51.3 | 41.8 | 39.6 |
| Released with other sources | 2 | 56.7 | 55.0 | 44.6 | 44.3 |
| Short ( min) | 8 | 68.3 | 64.8 | 57.2 | 53.6 |
| Medium (2–5 min) | 12 | 58.9 | 50.4 | 42.1 | 40.8 |
| Long ( min) | 7 | 53.0 | 52.1 | 40.3 | 38.4 |
| Property | Case A: knife retrieval | Case B: BWV understanding |
|---|---|---|
| Ground truth | Catalogue identity; objective and verifiable | Structured interpretation of an interaction; partly contestable |
| Output form | Ranked list of catalogue items | Free text with intervals |
| Metric maturity | Established retrieval metric (mAP) | Task metrics defined in this paper; support labelled by a judge validated against blinded human labels |
| Field validation | Deployed by UK forces; field accuracy not measured | None; experimental, not used by police; pilot benchmark on public footage only |
| Human role | Analyst confirms top match | Reviewer must verify each event used |
| Main error risk | Plausible wrong match under occlusion or open-set conditions | Fluent, unsupported or misattributed statements |
Appendix figures & tables9 assets
Supplementary material from the paper’s appendix.
Appendix
| # | Event type | What it covers | Videos |
|---|---|---|---|
| 1 | Armed suspect confrontation (edged or blunt weapon) | Officers face someone holding a knife, hammer or similar, with shouted commands and disarming | uk01, uk04, uk08, uk12, uk14, uk15, uk22, uk25 |
| 2 | Firearms incident | Armed police stops and discharges, or arrests of suspects linked to firearms | uk16, uk23 |
| 3 | Taser use | Taser drawn or fired to stop a suspect | uk01, uk14, uk24, uk25 |
| 4 | Assault on officers | An officer is attacked during the encounter | uk07, uk08, uk12, uk15 |
| 5 | Arrest at a live scene | Arrest straight after a violent incident, while the scene is still unfolding | uk01, uk04, uk10, uk14, uk22, uk25 |
| 6 | Planned arrest at a home | Officers arrive at an address, read the caution and detain the person | uk17, uk19, uk20 |
| Model | Videos | Sup. | Unsup. | Partial | Uncertain | Borderline | Duplicate | Total |
|---|---|---|---|---|---|---|---|---|
| Hybrid pipeline | 20 | 19 | 3 | 2 | 1 | 2 | 1 | 25 |
| Gemini 3.1 Pro | 21 | 19 | 4 | 1 | 1 | 2 | 1 | 25 |
| Qwen3.5-Omni-Plus | 19 | 18 | 4 | 1 | 2 | 2 | 0 | 25 |
| Nemotron 3 Nano Omni | 20 | 19 | 4 | 1 | 1 | 2 | 1 | 25 |
| Total | – | 75 | 15 | 5 | 5 | 8 | 3 | 100 |
| MSVD-CTN | MSRVTT-CTN | |||||
|---|---|---|---|---|---|---|
| Method | R-L | C | S | R-L | C | S |
| SEM-POS ( Nadeem et al., 2023 ) | 25.39 | 37.16 | 14.46 | 20.11 | 26.01 | 12.09 |
| AKGNN ( Hendria et al., 2023 ) | 25.11 | 35.08 | 14.55 | 21.42 | 25.90 | 11.99 |
| GIT ( Wang et al., 2022 ) | 27.51 | 45.63 | 15.58 | 24.51 | 32.43 | 13.70 |
| VideoLLaVA (zero-shot) ( Lin et al., 2024 ) | 21.80 | 30.55 | 14.67 | 19.33 | 16.24 | 12.49 |
| VideoLLaVA (LoRA FT) | 24.56 | 34.98 | 15.41 | 21.21 | 18.97 | 13.28 |
| Classification | Captioning | |||||||||
|---|---|---|---|---|---|---|---|---|---|---|
| Backbone | Integration | ESC50 | DCASE | VS | FSD | AS2M | AC | CL | AC | CL |
| Llama3.2-1B | PLITS | 64.45 | 37.69 | 51.57 | 25.23 | 9.08 | 0.59 | 0.34 | 16.30 | 10.96 |
| LAL | 76.70 | 40.97 | 60.87 | 31.44 | 11.83 | 0.66 | 0.38 | 16.97 | 11.87 | |
| PAL (LAL+LFST) | 87.40 | 46.23 | 56.03 | 43.91 | 14.74 | 0.72 | 0.42 | 18.08 | 12.58 | |
| Llama3.2-3B | PLITS | 70.40 | 40.62 | 61.40 | 28.88 | 10.84 | 0.63 | 0.35 | 16.81 | 11.35 |
| LAL | 82.15 | 43.21 | 65.78 | 34.29 | 12.91 | 0.67 | 0.38 | 17.80 | 12.18 | |
| Model | Size | Access | Input | Reference |
| NarrativeBridge | 80M | Local | V | ( Nadeem et al., 2025 ) |
| PAL + Llama 3.2 3B | 3B | Local | A | ( Alex et al., 2025 ) |
| GPT-5 post-processing (all systems) | n/d | API | Base-system text | ( OpenAI, 2025 ) |
| Hybrid pipeline (ours) | – | Hybrid local/API | V+A | This paper |
| Qwen3.5-Omni-Plus | n/d | API | V+A | ( Qwen Team, 2026 ) |
| Nemotron 3 Nano Omni | 30B-A3B | Open weights | V+A | ( NVIDIA, 2026 ) |
| ID | Force | Incident (de-identified) | Source | Year | Length |
|---|---|---|---|---|---|
| uk01 | Metropolitan Police | Arrest after a stabbing | Guardian News | 2026 | 0:54 |
| uk02 | Metropolitan Police | Armed response vehicle pursuits (compilation, partly body-worn) | The Independent | 2025 | 10:10 |
| uk03 | Metropolitan / Northamptonshire | Arrest of a woman; footage initially said not to exist | BBC London | 2026 | 8:38 |
| uk04 | Merseyside | Arrest of the perpetrator of a mass stabbing | Manchester Evening News | 2025 | 1:50 |
| uk05 | Hampshire | Handcuffing of a man at a stabbing scene | The Independent | 2026 | 2:44 |
| uk06 | West Midlands | Officers speak to a man at a murder scene | Daily Record | 2026 | 6:26 |
| UKBWV-27 | |
|---|---|
| Videos | 27 |
| Police forces | 16 |
| Published by police / news outlets | 3 / 24 |
| Upload years | 2017–2026 |
| Duration range | 0:26 to 15:44 |
| Total duration (min) | 118 |
| Backbone | Medium | Hard | Medium+D | Hard+D | Params |
|---|---|---|---|---|---|
| EfficientNet-B7 ( Tan and Le, 2019 ) | 79.4 | 56.1 | 73.6 | 45.5 | 70M |
| ConvNeXt-Base ( Liu et al., 2022b ) | 81.2 | 58.1 | 79.1 | 53.8 | 88M |
| SwinV2-Base ( Liu et al., 2022a ) | 82.4 | 59.2 | 80.2 | 54.4 | 88M |
| EVA02-Base ( Fang et al., 2024 ) | 83.1 | 61.2 | 81.5 | 57.9 | 88M |
| Configuration | Medium | Hard | Medium+D | Hard+D |
|---|---|---|---|---|
| Global average pooling | 81.1 | 58.2 | 79.1 | 53.1 |
| GeMNet | 83.1 | 61.2 | 81.5 | 57.9 |
| + Weibull pooling (ActNet) | 83.9 | 62.1 | 82.6 | 58.9 |
| + Local branch, orthogonal fusion | 85.6 | 64.1 | 84.7 | 62.1 |
| + Regional refinement (KHNet) | 86.5 | 65.3 | 85.7 | 63.4 |
Multimodal Model Evaluation
Multimodal model evaluation measures how reliably AI systems understand and generate content across text, images, audio, video, and other data.
See all 790 papers