Automatic visual regression testing is a challenging task when dealing with multimedia and dynamic content. Animations and randomly loaded content make this task impossible to achieve using pixel-wise comparison of screenshots.
Your work will be part of Adobe Experience Manager Platform (AEM) and will involve the development of machine learning – computer-vision models, that should be able to automatically detect high-level entities that make up a web page. Given a screenshot, the system will have to output a list of boxes with content type (picture, text, video, input) and other metadata that is useful for understanding how these components interact.
What you’ll do:
- Get up to speed with the world of enterprise-grade content management
- Gain hands-on experience with machine learning and computer vision
- Build a tool that generates the required annotation on a screenshot
- Showcase your work
What you need to succeed
- Analytical thinking and a desire to learn new things
- Knowledge of services-oriented architectures and server-side technologies
- Knowledge of Machine Learning principles and methods
- English communication skills, both written and verbal
- Linear algebra
- Deep learning toolkits: one of Tensorflow, Keras, Dynet, PyTorch
Anunț disponibil pentru studenții de la următoarele facultăți