Learning Generative Models of Images and Patterns

This Project is an continuation of our SIGGRAPH Asia 2017 paper on “Learning to Group Graphical Patterns”.
The paper introduced a novel deep learning approach for grouping discrete patterns common in graphical designs. The approach was based on a convolutional neural network architecture that learns a grouping measure defined over a pair of pattern elements. Motivated by perceptual grouping principles, the key feature of the network was the encoding of element shape, context, symmetries, and structural arrangements. These element properties are all jointly considered and appropriately weighted in the grouping measure. To better align the measure with the human perception of grouping, the network was trained on a large, human-annotated dataset of pattern groupings consisting of patterns at varying granularity levels, with rich element relations and varieties, tempered with noise and other data imperfections. TO BE CONT’D

Faculty Supervisor:

Richard Hao Zhang

Student:

Partner:

Discipline:

Computer science

Sector:

Technology; Information and Communications Technology; Education

University:

Simon Fraser University

Program:

Globalink Research Award