Project goals

The goal of this project is to develop a cloud platform for indexing large-scale images by the content using the technique of Big Data such as Hadoop, Spark, etc. The techniques of Deep Learning will be used for improving the accuracy of learning and retrieval phases. Within our platform, the user can provide the query image. The result will be represented by a set of similar images (Top-20 similar images). The platform can also index the query image if the user is not satisfied in order to improve precision.

Project leaders

  • Sidi Mahmoudi (
  • Mohammed Amin Belarbi (

Project organisation

During eNTERFACE’18 workshop, we would like to improve and develop three main tools for our platform :

Work Package 1: Deep learning features within Big Data
Work Package 2: GPU and cloud-based high-performance computing
Work package 3: Results validation