Timezone: »

Graph-based Isometry Invariant Representation Learning
Renata Khasanova · Pascal Frossard

Mon Aug 07 01:30 AM -- 05:00 AM (PDT) @ Gallery #58

Learning transformation invariant representations of visual data is an important problem in computer vision. Deep convolutional networks have demonstrated remarkable results for image and video classification tasks. However, they have achieved only limited success in the classification of images that undergo geometric transformations. In this work we present a novel Transformation Invariant Graph-based Network (TIGraNet), which learns graph-based features that are inherently invariant to isometric transformations such as rotation and translation of input images. In particular, images are represented as signals on graphs, which permits to replace classical convolution and pooling layers in deep networks with graph spectral convolution and dynamic graph pooling layers that together contribute to invariance to isometric transformation. Our experiments show high performance on rotated and translated images from the test set compared to classical architectures that are very sensitive to transformations in the data. The inherent invariance properties of our framework provide key advantages, such as increased resiliency to data variability and sustained performance with limited training sets.

Author Information

Renata Khasanova (Ecole Polytechnique Federale de Lausanne (EPFL))

Renata Khasanova is pursuing here PhD at the signal processing laboratory LTS4 at EPFL since 2014. She did her Bachelor and Master studies at Bauman Moscow State Technical University . She has also graduated from Yandex School of Data Analysis . Her scientific interests include Deep learning and Graph Signal Processing.

Pascal Frossard (EPFL)

Related Events (a corresponding poster, oral, or spotlight)

More from the Same Authors