AI lounge: Automated Acoustic Data Processing with Deep Learning

Welcome to this new AI Lounge. We are happy to present a new interesting presentation.

Poster about the AI lounge (pdf).

Coffee, the and cookies will be served. 

Looking forward to seeing you at the last AI lounge before the summer break!

Automated Acoustic Data Processing with Deep Learning

by  Olav Brautaset is a researcher at Norsk Regnesentral (Norwegian Computing Center) in Oslo, an independent research institute in applied statistical modelling and machine learning. He is part of the image analysis group, working on problems in seismic, marine and medical image data.

We are currently developing a deep learning model to estimate the amount of different fish species in acoustic echo sounder images. The model is based on UNet – a convolutional neural network architecture. This produces a segmentation of the input image, classifying each pixel to either a fish species or background. The amount of fish in an image can then be obtained by summing the intensities of all pixels classified as a particular species from the segmentation.

Questions

 

Organizer

Andrea Gasparini, Anne Schad Bergsaker and Thomas R?blitz
Tags: AI, machine learning, Norsk Regnesentral, deep learning, USIT, ITF, UB, AI lounge
Published May 20, 2019 9:15 AM - Last modified May 20, 2019 9:33 AM