Computer vision methods are quite advanced nowadays. They are often used in industry, factories, but wider usage of computer vision is partly limited by personal data protection, GDPR and other ethical questions. Therefore the application of computer vision in process automation sometimes is limited and not fully utilized. Building new solutions, which take into account all possible prescriptions and computer vision possibilities could boost the automation level and contribute to the technological development. Therefore it is important to find the best cases for computer vision usage, its practical implementation and its impact on the automated process efficiency from different perspectives.
The direction for the challenge ideas, which could be converted into multiple final thesis or individual project topics is the following:
- Automatic logging, marking of student’s participation in the lecture based on computer vision.
- Sentiment analysis in images based on object detection methods.
- A location determination method based on photos and metadata provided on social networks.
- Color reconstruction of Čiurlionis’ paintings from black and white photos using adversarial networks
- Optical note recognition and writer identification in Čiurlionis’ piano music manuscripts
- Handwritten music note recognition for piano manuscripts using deep neural networks
Regarding more details on the challenge and topics contact Pavel Stefanovič (topics 1-3) and Jūratė Janutėnaitė Bogdanienė (topics 4-6), VILNIUS TECH.