Tereza’s internship at Fraunhofer IWU

Date: 18. 6. - 17. 9. 2021

“I have been on three-months internship at Fraunhofer institute IWU in Chemnitz where I became a part of several IWU research projects. The internship took place under the guidance of my supervisor Alexander Pierer.

First, I created a tutorial about accelerating machine learning (ML) using an ONNX that is wanted to incorporate into IWU ML applications. The tutorial included a theoretical summary of this issue, description of the individual parts, possible ways of working with ONNX format, implementation of ONNX format, ONNX training and model creation, working with ONNX Runtime framework and sample applications using Azure service that support ONNX.

Another part of the internship was to create a module for defect detection in textiles. For this purpose, ONNX was used to train and deploy the created model into C#, which was then implemented into the Xeidana framework.

The last part was the creation of a program for defect detection on railways. That part is still in process, especially an implementation of different ML approaches whose performances will be compared in the resulting paper.

The internship was very beneficial, I gained new knowledge in the field of machine learning, specifically in the area of machine vision and unsupervised ML approaches.”

Tereza in front of Fraunhofer IWU and her workplace at the Frauhofer IWU.