# H4VDM: H.264 Video Device Matching

arXiv: https://arxiv.org/abs/2210.11549 · HTML: https://mlsanity.com/paper/2210.11549/

## Key facts

- No citing papers in the map yet (data through 14 Oct 2026).
- Builds on 18 of its 56 references that are in the map.
- First posted on arXiv 20 Oct 2022.

## Builds on (most cited first, 18 of 18)

- [Attention Is All You Need](https://mlsanity.com/paper/1706.03762/) (2017)
- [Deep Residual Learning for Image Recognition](https://mlsanity.com/paper/1512.03385/) (2015)
- [Adam: A Method for Stochastic Optimization](https://mlsanity.com/paper/1412.6980/) (2014)
- [Language Models are Few-Shot Learners](https://mlsanity.com/paper/2005.14165/) (2020)
- [An Image is Worth 16x16 Words: Transformers for Image Recognition at Scale](https://mlsanity.com/paper/2010.11929/) (2020)
- [Swin Transformer: Hierarchical Vision Transformer using Shifted Windows](https://mlsanity.com/paper/2103.14030/) (2021)
- [Efficient Estimation of Word Representations in Vector Space](https://mlsanity.com/paper/1301.3781/) (2013)
- [On the Properties of Neural Machine Translation: Encoder-Decoder Approaches](https://mlsanity.com/paper/1409.1259/) (2014)
- [Video Swin Transformer](https://mlsanity.com/paper/2106.13230/) (2021)
- [VideoBERT: A Joint Model for Video and Language Representation Learning](https://mlsanity.com/paper/1904.01766/) (2019)
- [Audio Transformers](https://mlsanity.com/paper/2105.00335/) (2021)
- [Forensic Similarity for Digital Images](https://mlsanity.com/paper/1902.04684/) (2019)
- [An Overview of Recent Work in Media Forensics: Methods and Threats](https://mlsanity.com/paper/2204.12067/) (2022)
- [CNN-based fast source device identification](https://mlsanity.com/paper/2001.11847/) (2020)
- [Facing Device Attribution Problem for Stabilized Video Sequences](https://mlsanity.com/paper/1811.01820/) (2018)
- [Video Camera Identification from Sensor Pattern Noise with a Constrained ConvNet](https://mlsanity.com/paper/2012.06277/) (2020)
- [Forensic Analysis of Video Files Using Metadata](https://mlsanity.com/paper/2105.06361/) (2021)
- [Mitigation of H.264 and H.265 Video Compression for Reliable PRNU Estimation](https://mlsanity.com/paper/1905.09611/) (2019)

## Abstract (arXiv)

Methods that can determine if two given video sequences are captured by the same device (e.g., mobile telephone or digital camera) can be used in many forensics tasks. In this paper we refer to this as "video device matching". In open-set video forensics scenarios it is easier to determine if two video sequences were captured with the same device than identifying the specific device. In this paper, we propose a technique for open-set video device matching. Given two H.264 compressed video sequences, our method can determine if they are captured by the same device, even if our method has never encountered the device in training. We denote our proposed technique as H.264 Video Device Matching (H4VDM). H4VDM uses H.264 compression information extracted from video sequences to make decisions. It is more robust against artifacts that alter camera sensor fingerprints, and it can be used to analyze relatively small fragments of the H.264 sequence. We trained and tested our method on a publicly available video forensics dataset consisting of 35 devices, where our proposed method demonstrated good performance.