Project

Improving deep learning for electron microscopy

Publisher

Supervisor

Location

Greater Copenhagen area

Deep learning is the application of neural networks with many layers to machine learning problems. We expect that deep learning can be a game-changer for analyzing data from Transmission Electron Microscopy (TEM). It can automatize the analysis of large amounts of data, eliminate the bias of the person analysing the images, provide quantitative and statistical data, and even help extracting information that is normally difficult to get from TEM.

In a collaboration between the CAMD theory group at DTU Physics and the electron microscopists at DTU CEN/Danchip we have developed a method for training deep convolutional neural networks on simulated TEM images and applying them on real experimental data [1].  We have made our prototype software publicly available [2].

Unlike most image analysis problems, we are in the beneficial and quite unusual situation of having access to almost free training data, since high-resolution TEM images can be simulated reliably from atomic structures, including realistic noise and imperfections in the microscope optics.

The neural network architecture is inspired by the networks generally used for automated image analysis and segmentation.  However, the network architecture has almost not been optimized, nor has the training procedure.  We expect that the performance can be significantly enhanced by addressing this.

This project has two aspects, a BSc project will focus one one of them while a MSc will focus on both.

* To improve the neural network by optimizing the network architecture and the training algorithm, and by incorporating recent ideas from the litterature on image analysis and image segmentation.

* To turn the current collection of scripts and software snippets into a package that can realistically be used by an electron microscopist with no prior experience in machine learning.experience with 



Figure: The architecture of the neural network [1]. Information flows from left to right. The different-colored rectangles refer to the different architecture elements. Below the rectangles, the spatial and channel dimensions are given as height × width × number of channels (feature maps). The features are downsampled in an encoding path and upsampled through a decoding path in order to represent non-local information. Skip connections ensure that it is possible to retain the original spatial information. Although the size of the input images is shown as 256 × 256, this is not part of the network architecture, and the network can be used on images of any size.


References:

[1] J. Madsen et al.: A Deep Learning Approach to Identify Local Structures in Atomic-Resolution Transmission Electron Microscopy Images,  Adv. Theory Simul. 1, 1800037 (2018).  Preprint available at https://arxiv.org/abs/1802.03008


Requirements

Prior knowledge of Python and neural networks. Some knowledge of solid state physics and/or electron microscopy is an advantage, but not required.

Search in postings
Contact

Company / Organization

DTU Fysik

Name

Jakob Schiøtz

Position

Professor

Mail

schiotz@fysik.dtu.dk

Supervisor info

Artificial Intelligence and Data

Supervisor

Jakob Schiøtz

Co-supervisors

Thomas Willum Hansen

ECTS credits

15 - 35

Type

BSc project, MSc thesis

Bachelor in General Engineering

Supervisor

Jakob Schiøtz

Co-supervisors

Thomas Willum Hansen

ECTS credits

15 - 35

Type

BSc project, MSc thesis

BSc in Earth and Space Physics and Engineering

Supervisor

Jakob Schiøtz

Co-supervisors

Thomas Willum Hansen

ECTS credits

15 - 35

Type

BSc project, MSc thesis

BSc in Mathematics and Technology

Supervisor

Jakob Schiøtz

Co-supervisors

Thomas Willum Hansen

ECTS credits

15 - 35

Type

BSc project, MSc thesis

BSc in Physics and Nanotechnology

Supervisor

Jakob Schiøtz

Co-supervisors

Thomas Willum Hansen

ECTS credits

15 - 35

Type

BSc project, MSc thesis

MSc Eng in Applied Chemistry

Supervisor

Jakob Schiøtz

Co-supervisors

Thomas Willum Hansen

ECTS credits

15 - 35

Type

BSc project, MSc thesis

MSc in Computer Science and Engineering

Supervisor

Jakob Schiøtz

Co-supervisors

Thomas Willum Hansen

ECTS credits

15 - 35

Type

BSc project, MSc thesis

Msc in Earth and Space Physics and Engineering

Supervisor

Jakob Schiøtz

Co-supervisors

Thomas Willum Hansen

ECTS credits

15 - 35

Type

BSc project, MSc thesis

MSc in Electrical Engineering

Supervisor

Jakob Schiøtz

Co-supervisors

Thomas Willum Hansen

ECTS credits

15 - 35

Type

BSc project, MSc thesis

MSc in Mathematical Modelling and Computation

Supervisor

Jakob Schiøtz

Co-supervisors

Thomas Willum Hansen

ECTS credits

15 - 35

Type

BSc project, MSc thesis

MSc in Physics and Nanotechnology

Supervisor

Jakob Schiøtz

Co-supervisors

Thomas Willum Hansen

ECTS credits

15 - 35

Type

BSc project, MSc thesis

Technical University of Denmark

For almost two centuries DTU, Technical University of Denmark, has been dedicated to fulfilling the vision of H.C. Ørsted – the father of electromagnetism – who founded the university in 1829 to develop and create value using the natural sciences and the technical sciences to benefit society.


Today, DTU is ranked as one of the foremost technical universities in Europe, continues to set new records in the number of publications, and persistently increases and develops our partnerships with industry, and assignments accomplished by DTU’s public sector consultancy.

Find us here

Anker Engelunds Vej 1
Bygning 101A
2800 Kgs. Lyngby

Denmark



Tlf. (+45) 45 25 25 25

CVR-nr. 30 06 09 46

All vacant positions
 

loading..