Research Technician in computer vision / deep learning for fighting against human trafficking for the European HEROES Project Contract

Contexte et atouts du poste

The Inria STARS team is seeking a young backend developer with a background in computer vision (4 months). We are especially interested in candidates with backgrounds in computer vision, deep learning, machine learning, and applied mathematics.

The candidate is expected to conduct object detection/tracking research, specifically in videos for human trafficking for the HEROES Project – Novel Strategies to Fight Child Sexual Exploitation and Human Trafficking Crimes and Protect their Victims. 

Trafficking of human beings (THB) and child sexual abuse and exploitation (CSA/CSE) are two big problems in our society. Inadvertently, new information and communication technologies (ICTs) have provided a space for these problems to develop and take new forms, made worse by the lockdown caused by the COVID-19 pandemic. At the same time, technical and legal tools available to stakeholders that prevent, investigate, and assist victims – such as law enforcement agencies (LEAs), prosecutors, judges, and civil society organizations (CSOs) – fail to keep up with the pace at which criminals use new technologies to continue their abhorrent acts. The HEROES project’s main objective is to use technology to improve the way in which help and support can be provided to victims of THB and CSA/CSE.

The partners include Inria and Universidade de Brasília – UnB, International Center for Missing and Exploited Children, Secretaria de Inteligencia Estratégica de Estado – Presidencia. INRIA Grant is 200 Keuros out of 3 Meuros.

For more information, see The recruitee will mainly work with Francois Bremond, but beyond that, he/she will be able to collaborate with members of the STARS team. Support for traveling e.g. to conferences or other scientific meetings is provided.

 —————————

The last decade has seen the emergence of machine learning methods for decision support and information retrieval in large volumes of data, such as Boosting, Support Vector Machines, Random Forest and artificial neural networks known as “deep convolutional architectures”. These deep learning techniques provide solutions for complex applications such as information retrieval from huge volumes of multimedia content, speech recognition, social community extraction from social networks or knowledge extraction from textual data. In particular, CNN algorithms for deep convolutional neural networks have been applied to images and videos, for people monitoring applications such as posture detection and estimation, gesture recognition and action detection algorithms.

The STARS group works on the automatic interpretation of images and video sequences using the methods described above. These interpretations include the automatic detection of CSAM (Children Sexual Abuse Material) from videos or images. Not having access to this type of data, the solution proposed by the STARS team is to couple the detection of minors using age estimation algorithms with an explicit content recognition module. In order to provide the most accurate age estimation possible, specific algorithms have been developed in the laboratory, based in particular on spatio-temporal body information, or on facial ageing features with a system of comparison with references.

Mission confiée

Tasks involved:

In this context, and as part of the HEROES project ( the subject of this mission is to integrate the various models trained, notably explicit content detection and age estimation, into an application with an intuitive graphical interface to enable users (police forces) to evaluate these algorithms on their data sets and captured material. This application is part of the demonstrator to be provided during the project pilot in the form of a web-based application (hosted in a closed network), enabling more efficient deployment within police infrastructures, and must integrate writing to and reading from a database in order to retain the results of previously analyzed files.

Principales activités

The Inria STARS team is seeking a Post Doctoral researcher with a strong computer vision and mathematics background. We are especially interested in candidates with solid computer vision, deep learning, machine learning, and applied mathematics backgrounds.

Team

The STARS research team combines advanced theory with a cutting-edge practice focusing on cognitive vision systems.

Team website

Compétences

Avantages

Rémunération

From 1810 € gross monthly (according to degree and experience)