Advanced search
Start date

Image source camera attribution

Full text
Filipe de Oliveira Costa
Total Authors: 1
Document type: Master's Dissertation
Press: Campinas, SP.
Institution: Universidade Estadual de Campinas (UNICAMP). Instituto de Computação
Defense date:
Examining board members:
Anderson de Rezende Rocha; Marco Antônio Piteri; Ricardo da Silva Torres
Advisor: Anderson de Rezende Rocha

Image's integrity and authenticity verification is paramount when it comes to a court of law. Just like we do in ballistics tests when we match a gun to its bullets, we can identify a given digital camera that acquired an image under investigation. In this work, we discussed approaches for identifying whether or not a given image under investigation was captured by a specific digital camera. We carried out the research under two vantage points: (1) verification, in which we are interested in verifying whether or not a given camera captured an image under investigation; and (2) recognition, in which we want to verify if an image was captured by a given camera (if any) from a pool of devices, and to point out such a camera. We performed this investigation considering an open set scenario, under which we can not rely on the assumption of full access to all of the investigated devices. We also tried to solve the device linking problem, where we aim at verifying if an image pair was generated by the same camera, without any information about the source of images. Our approaches reported good results, in terms of being capable of identifying the specific device that captured a given image including its model, brand, and even serial number (AU)

FAPESP's process: 11/03808-3 - Digital images source attribution
Grantee:Filipe de Oliveira Costa
Support Opportunities: Scholarships in Brazil - Master