Topic

Computer vision

Computer vision work on this site is research rather than product. The MSc dissertation asked whether the iterative evolution of a level set segmentation, a method normally driven by a partial differential equation, can be learned as a sequence by recurrent networks, and compared an echo state network against trained convolutional RNN, LSTM, GRU and 3D CNN models on four image datasets. Earlier, a Wits honours project team, his included, built a TurtleBot that follows an object using Python and ROS.

1 research · 1 publications · 1 talks and sessions · Wikidata Q844240

Research

Echo state networks for level set image segmentationMSc research at Wits: can recurrent models learn the iterative evolution of a level set segmentation? Echo state networks against trained RNN, LSTM and GRU.

Publications

Learning Level Set Method by Echo State Network for Image SegmentationStudies echo state networks as a cheaper alternative to recurrent networks trained by backpropagation, applied to learning variational level set image segmentation as a spatiotemporal, data driven method. Five convolutional architectures were compared (ESN, RNN, GRU, LSTM and a 3D CNN) on four datasets; the GRU and LSTM variants performed best. The ESN performed poorly, which the dissertation attributes largely to the reservoir's leaking rate and spectral radius.

Talks and sessions

Follow me Turtle Bot ProjectOur follow me turtle bot project, used python and ROS to design turtle-bot that follows an object. This was done at the Witwatersrand University school of Computer Science and Applied Mathematics.

Subjects that share work with this one. Those with their own page are linked.

Artificial intelligence
Echo state networks
Image segmentation
Level set method
Recurrent neural networks
Reservoir computing
Deep learning
Robotics

Research

Publications

Talks and sessions

Related topics