Dr. Lázaro Janier Gonzalez-Soler

Project management of the European research project ‘Einstein’ for developing computer vision algorithms against the detection of identity fraud. Research and development of computer vision solutions for identity fraud detection in embedded systems. Generation of morphing attack databases using latent diffusion models. Implementation of zero-, one- and few-shot inference strategies on vision language models against identity fraud attacks.

Research and development of biometric technologies using hybrid-conditional latent diffusion models for the synthesis of large synthetic biometric databases. Optimisation of Deep Neural Networks for hand recognition and image quality estimation. Finetuning cutting-edge semantic segmentation algorithms using synthetic data for tattoo segmentation. Supervision of a master's thesis focused on the development of multimodal AI methods for the synthesis of biometric data.

Research and development of machine learning algorithms for identity fraud detection. Definition, training and finetuning of DNN architectures for fingerprint, face, and voice presentation attack detection (PAD). Conceptualisation and implementation of DNN-based anomaly detection algorithms against facial morphing attacks and presentation attacks.

Development of data collection pipelines using web scraping technologies to process more than 500,000 data points per week.

Development of algorithms based on abstract syntax trees to automatically interpret ASTM, Kermit and HL7 message protocols from different laboratory instrument vendors.

GPU optimisation of overlap clustering algorithms accelerating social network analysis by 175x. Optimisation and upgrade of a GPU-based fingerprint identification system for Jetson TK1, improving fingerprint matching efficiency by up to 50%.