2021-11-25 on Morphing Attack Detection

da/sec scientific talk on Morphing Attack Detection

Topic: Morphing Detection using Siameses networks and few-shots learning approach

by Juan Tapia
online Big Blue Button Room: D19/2.03a, November 25, 2021 (Thursday), 12.00 noon

Keywords — Morhing, Siameses network, few shots


Siamese neural network based on a few-shot learning model tries to deal with morphing images issue with few labelled data.
In particular, a Siamese convolutional neural network (CNN) is constructed to improve high-dimensional feature learning and further facilitate the identification of novel classes based on an optimized feature representation.
Few-shot learning is an emerging type of transfer learning technique. By reusing the transferable knowledge of existing models, a classifier can be built to identify the novel category using only a few labelled training samples.