Título: TARP Fingerprinting: a countermeasure resistant HTML5 Canvas based browser Fingerprinting mechanism

Horário: 14:00

Data: 27/11/2019

Local: GREat - Bloco 942-A (Sala de Seminários)

Resumo:

This work presents a novel HTML5 Canvas based browser fingerprinting mechanism which is resistant to countermeasures. More specifically, we propose a variation of the original Canvas Fingerprinting technique that is extended to identify client-side <canvas> element tampering. Because of that, the mechanism discussed in this work is able to identify and avoid systems and techniques used as countermeasures to traditional Canvas Fingerprinting, specifically the ones that are based on blocking or modifying the <canvas> element normal behavior. The proposed mechanism reaches this goal through an entropy modulation strategy with regards to HTML5 canvas based signatures in a way that not only works for fingerprinting (high entropy) but also for tampering identification (low entropy). Furthermore, we present one of the only two large-scale (64.000+ samples) Canvas fingerprinting entropy measuring studies found in the literature to this date.

Banca:

• Prof. Dr. Miguel Franklin de Castro(MDCC/UFC - Orientador)
• Prof. Dr. Fernando Antônio de Carvalho Gomes (UFC - Coorientador)
• Prof. Dr. André Luiz Moura dos Santos (UECE)