The analysis of fingerprints continues to be one of the most relevant tools in criminal investigations, but it is also a complex process that requires great precision and expertise. In this context, the National Institute of Standards and Technology (NIST) in the United States has taken an important step to modernise this field with the launch of new resources that combine big data and open-source technology.

NIST has recently completed its Special Database 302 (SD 302), a set that includes approximately 10,000 fingerprints collected in controlled environments from 200 volunteers. Although this database has existed since 2019, until now only a portion of the images had detailed annotations. With this new update, all fingerprints have been completely annotated, significantly increasing their value for research and training.

These annotations are especially relevant because they indicate the quality of different areas of each imprint using colour codes. In practice, this allows for easier identification of which parts of a fingerprint contain useful information for identification and which may be less reliable. This distinction is fundamental for both human examiners and automated systems, as the fingerprints collected at crime scenes are often incomplete, blurred, or partially deteriorated.

From the perspective of security and forensic investigation, the availability of a fully annotated dataset represents a key advancement. It not only facilitates the training of new professionals, but also allows for the development and validation of artificial intelligence algorithms with a solid empirical foundation. At a time when AI plays an increasing role in the analysis of digital evidence, having quality data is essential to ensure reliable and reproducible results.

At the same time, NIST has released a new open-source software called OpenLQM, an evolution of a tool previously used by law enforcement in the United States. This software is designed to automatically assess the quality of fingerprints. Its operation is relatively simple: it analyses an image and assigns a score between 0 and 100 that reflects the level of detail and usefulness of the imprint.

This functionality has very important practical implications. In a real investigation, analysts may encounter hundreds of fingerprints collected at the same scene. Automatically classifying them according to their quality allows to prioritise those that are more likely to lead to a positive identification, thus reducing the time and resources required. Moreover, the use of an objective metric helps improve consistency among different examiners, a critical aspect in judicial processes.

Another notable element is that OpenLQM is open-source and compatible with multiple operating systems, including Windows, Mac, and Linux. This facilitates its global adoption by both law enforcement agencies and research centres and universities. This openness also fosters transparency and collaboration, two values that are becoming increasingly important in the development of technologies applied to security.

Overall, the combination of the fully annotated SD 302 dataset and the Open LQM software provides a powerful platform for advancing the science of fingerprint identification. These resources not only enhance the tools currently available, but also lay the groundwork for future innovations in the field of digital forensics and security.

Ultimately, the NIST initiative exemplifies how the integration of quality data, open tools, and artificial intelligence can transform traditional processes and enhance their efficiency and reliability. For security professionals, these advancements represent a clear opportunity to enhance analytical capabilities and adapt to an increasingly technological and demanding environment.

_____

Aquest apunt en català / Esta entrada en español / Post en français