# Using Wavelet Domain Fingerprints to Improve Source Camera Identification

arXiv: https://arxiv.org/abs/2507.01712 · HTML: https://mlsanity.com/paper/2507.01712/

## Key facts

- No citing papers in the map yet (data through 14 Oct 2026).
- Builds on 4 of its 98 references that are in the map.
- First posted on arXiv 2 Jul 2025.

## Builds on (most cited first, 4 of 4)

- [Deep learning for source camera identification on mobile devices](https://mlsanity.com/paper/1710.01257/) (2017)
- [CNN-based fast source device identification](https://mlsanity.com/paper/2001.11847/) (2020)
- [Beyond PRNU: Learning Robust Device-Specific Fingerprint for Source Camera Identification](https://mlsanity.com/paper/2111.02144/) (2021)
- [Cyber-Security Internals of a Skoda Octavia vRS: A Hands on Approach](https://mlsanity.com/paper/1910.09410/) (2019)

## Abstract (arXiv)

Camera fingerprint detection plays a crucial role in source identification and image forensics, with wavelet denoising approaches proving particularly effective for extracting sensor pattern noise (SPN). In this article, we introduce the concept of a wavelet domain (WD) fingerprint, redefining the representation of the extracted fingerprint from the conventional image domain to the native wavelet coefficient domain. Rather than reconstructing the fingerprint as a spatial domain image, fingerprint comparison is performed directly on the wavelet coefficients, eliminating the final inverse transform and subsequent image-domain post-processing. This reformulation streamlines the fingerprint extraction and comparison pipeline while preserving the information required for source camera identification. The proposed framework is applicable to existing wavelet-based SPN extraction methods and is demonstrated using two representative state-of-the-art pipelines. Experimental results on real-world datasets show that the proposed approach significantly reduces computational cost, making it well-suited for large-scale source camera identification applications.