Nowadays, computer vision is becoming increasingly popular both in the industry and in various research laboratories. The applications of this field are nearly endless, ranging from self-driving cars to automated production lines and security technologies. Within the field of computer vision, in the past one to two decades, various learning methods — particularly the application of deep neural networks — promise the greatest opportunities.

In the project laboratory topic, the student’s task is to become familiar with deep learning methods and their applications, and then use them to provide a solution to a selected problem.

Within this field, the following tasks can be performed:

• Classification of visual (or possibly audio) information. (Recognizing people, animals, vehicle types, tumors, words, intentions, actions, traffic signs, etc.)
• Detection/segmentation performed on visual (or possibly audio) information. (Identifying relevant objects for self-driving cars, speech enhancement, etc.)
• Tasks performed on sequences of visual (or possibly text) data. (Video analysis, image captioning, speech recognition, etc.)
• Generation of images (or possibly sounds).
• Development of higher cognitive functions based on audiovisual input. (Playing various computer games, implementing attention and memory, question-answer systems.)
• Investigation of high-level modern learning methods. (Curiosity-driven learning, meta-learning, etc.)
• Individual ideas.

Requirements for the topic:

• Knowledge of object-oriented programming (C++/Python, or similar).
• Good command of English.

Recommended for the topic:

• Minimal image (signal) processing knowledge.
• Minimal mathematical knowledge.