Register now After registration you will be able to apply for this opportunity online.
This opportunity is not published. No applications will be accepted.
GoodOnFocus: Computer Vision Thesis
In supermarkets, grocery stores or warehouses, we often
search products and goods which are suitable to our needs:
e.g. during shopping, we want to quickly know the ingredients
of a product. Today, consumer need to read small print on a
product’s backside or need to scan a product’s barcode – a
process which requires sufficient illumination, high-resolution
camera, proximity between product and smartphone, and an
active command from the user or app to scan the barcode.
Soon, this process might become more natural – e.g., through
smart glasses or Mixed Reality headsets, which automatically
detect product category or size, even without scanning its barcode.
In “GoodOnFocus” project, you will use computer
vision to automatically identify products.
Our current database contains approximately 15000 products,
their image, size, description, nutrient, and category data. We
constantly extend the database and use it for training and test
purposes. You will develop a machine learning script that predicts
the category of products (e.g., sweetened beverage) from
their images and calculates the confidence of that prediction.
Therefore, you can build your project upon recent and relevant
literature and use existing machine learning libraries such as
Keras, PyTorch or Tensorflow.
Our current database contains approximately 15000 products, their image, size, description, nutrient, and category data. We constantly extend the database and use it for training and test purposes. You will develop a machine learning script that predicts the category of products (e.g., sweetened beverage) from their images and calculates the confidence of that prediction. Therefore, you can build your project upon recent and relevant literature and use existing machine learning libraries such as Keras, PyTorch or Tensorflow.
Dr. Kenan Bektas: kenan.bektas@unisg.ch
Dr. Klaus Fuchs: fuchsk@ethz.ch
Dr. Kenan Bektas: kenan.bektas@unisg.ch Dr. Klaus Fuchs: fuchsk@ethz.ch