CCTV-view classroom incident recognition benchmark (synthetic plus real-world preschool pose data)¶
Product or Programme · All products and programmes
Evidence · 3 claims (3 for) · 1 study (1 design),q2· 0 of 1 report an effect size · 3 claims rest on one study
Description¶
A privacy-aware benchmark dataset from Paritosh Parmar et al. combining synthetic CCTV-style videos with anonymized real-world preschool pose data annotated for classroom incident-recognition research.
Components¶
Claims¶
- A distilled multi-order kinematic model outperforms pose-based baselines on a synthetic classroom incident dataset while using one-tenth as many parameters as PoseC3D [+M]
- Models trained on synthetic classroom incident data transfer zero-shot to real-world footage, with the proposed method best at 63.41% though a synthetic-to-real domain gap remains [+M]
- Classroom incident recognition is substantially harder than established general action-recognition benchmarks, with PoseC3D dropping from over 94% accuracy to 70.54% [+W]
Related Products and Programmes¶
-
Key Sources¶
- Paritosh Parmar, Landy Lan, Hong Yang, Chen Yi, Chiat Pin Tay. (2026). Robust and Efficient Motion Reasoning for Privacy-Aware Classroom Incident Recognition. arXiv preprint. https://arxiv.org/abs/2608.05115
Page metadata
| Field | Value |
|---|---|
type |
product |
id |
cctv-view-classroom-incident-recognition-benchmark-synthetic-plus-real-world-preschool-pos |
title |
CCTV-view classroom incident recognition benchmark (synthetic plus real-world preschool pose data) |
description |
A privacy-aware benchmark dataset from Paritosh Parmar et al. |
status |
review |
product_kind |
dataset |
generated |
by process:settle-candidates on 2026-10-11 |
trust |
unverified — no one has confirmed this page's content |
Frontmatter is the page's machine-readable record — see the Schema & Guide.