# PrivHAR-Bench: A Graduated Privacy Benchmark Dataset for Video-Based Action Recognition

arXiv: https://arxiv.org/abs/2604.00761 · HTML: https://mlsanity.com/paper/2604.00761/

## Key facts

- No citing papers in the map yet (data through 14 Oct 2026).
- Builds on 24 of its 53 references that are in the map.
- First posted on arXiv 1 Apr 2026.

## Builds on (most cited first, 24 of 24)

- [Deep Residual Learning for Image Recognition](https://mlsanity.com/paper/1512.03385/) (2015)
- [Quo Vadis, Action Recognition? A New Model and the Kinetics Dataset](https://mlsanity.com/paper/1705.07750/) (2017)
- [The Kinetics Human Action Video Dataset](https://mlsanity.com/paper/1705.06950/) (2017)
- [SlowFast Networks for Video Recognition](https://mlsanity.com/paper/1812.03982/) (2018)
- [Is Space-Time Attention All You Need for Video Understanding?](https://mlsanity.com/paper/2102.05095/) (2021)
- [Datasheets for Datasets](https://mlsanity.com/paper/1803.09010/) (2018)
- [VideoMAE: Masked Autoencoders are Data-Efficient Learners for Self-Supervised Video Pre-Training](https://mlsanity.com/paper/2203.12602/) (2022)
- [Image Super-Resolution Using Deep Convolutional Networks](https://mlsanity.com/paper/1501.00092/) (2014)
- [Spatial Temporal Graph Convolutional Networks for Skeleton-Based Action Recognition](https://mlsanity.com/paper/1801.07455/) (2018)
- [A Closer Look at Spatiotemporal Convolutions for Action Recognition](https://mlsanity.com/paper/1711.11248/) (2017)
- [Two-Stream Convolutional Networks for Action Recognition in Videos](https://mlsanity.com/paper/1406.2199/) (2014)
- [NTU RGB+D: A Large Scale Dataset for 3D Human Activity Analysis](https://mlsanity.com/paper/1604.02808/) (2016)
- [NTU RGB+D 120: A Large-Scale Benchmark for 3D Human Activity Understanding](https://mlsanity.com/paper/1905.04757/) (2019)
- [GaitSet: Regarding Gait as a Set for Cross-View Gait Recognition](https://mlsanity.com/paper/1811.06186/) (2018)
- [Towards Privacy-Preserving Visual Recognition via Adversarial Training: A Pilot Study](https://mlsanity.com/paper/1807.08379/) (2018)
- [STPrivacy: Spatio-Temporal Privacy-Preserving Action Recognition](https://mlsanity.com/paper/2301.03046/) (2023)
- [Bullying10K: A Large-Scale Neuromorphic Dataset towards Privacy-Preserving Bullying Recognition](https://mlsanity.com/paper/2306.11546/) (2023)
- [PrivHAR: Recognizing Human Actions From Privacy-preserving Lens](https://mlsanity.com/paper/2206.03891/) (2022)
- [Privacy-Preserving Action Recognition via Motion Difference Quantization](https://mlsanity.com/paper/2208.02459/) (2022)
- [Privacy Assessment on Reconstructed Images: Are Existing Evaluation Metrics Faithful to Human Perception?](https://mlsanity.com/paper/2309.13038/) (2023)
- [Visual Content Privacy Protection: A Survey](https://mlsanity.com/paper/2303.16552/) (2023)
- [Can masking background and object reduce static bias for zero-shot action recognition?](https://mlsanity.com/paper/2501.12681/) (2025)
- [Video-DPRP: A Differentially Private Approach for Visual Privacy-Preserving Video Human Activity Recognition](https://mlsanity.com/paper/2503.02132/) (2025)
- [Seeing Beyond the Scene: Analyzing and Mitigating Background Bias in Action Recognition](https://mlsanity.com/paper/2512.17953/) (2025)

## Abstract (arXiv)

Existing research on privacy-preserving Human Activity Recognition (HAR) typically evaluates methods against a binary paradigm: clear video versus a single privacy transformation. This limits cross-method comparability and obscures the nuanced relationship between privacy strength and recognition utility. We introduce textitPrivHAR-Bench, a multi-tier benchmark dataset designed to standardize the evaluation of the textitPrivacy-Utility Trade-off in video-based action recognition. PrivHAR-Bench applies a graduated spectrum of visual privacy transformations: from lightweight spatial obfuscation to cryptographic block permutation, to a curated subset of 15 activity classes selected for human articulation diversity. Each of the 1,932 source videos is distributed across 9 parallel tiers of increasing privacy strength, with additional background-removed variants to isolate the contribution of human motion features from contextual scene bias. We provide lossless frame sequences, per-frame bounding boxes, estimated pose keypoints with joint-level confidence scores, standardized group-based train/test splits, and an evaluation toolkit computing recognition accuracy and privacy metrics. Empirical validation using R3D-18 demonstrates a measurable and interpretable degradation curve across tiers, with within-tier accuracy declining from 88.8% (clear) to 53.5% (encrypted, background-removed) and cross-domain accuracy collapsing to 4.8%, establishing PrivHAR-Bench as a controlled benchmark for comparing privacy-preserving HAR methods under standardized conditions. The dataset, generation pipeline, and evaluation code are publicly available.