Hi there ! The project goal was to create a tool, which eventually will aid designers during their brainstorming process. By choosing two distinct objects (represented as nodes on a weighted graph), characterized by different categories, users can find the shortest path between the two, gradually through other objects sharing similarities. The shortest Pathfinding process was implemented in two different ways - using Dijkstra and A-star algorithms. Finally, for each node in the shortest path, the top results images, matching its characteristics on ‘Pinterest’ will be displayed – building a visual map for the user.
project DEMO can be seen here: https://drive.google.com/open?id=1H6gIFFglyUADAbtSACEvFcRkv26k3AQM
Course: AI and ML for interactive products.
Guiding: Michael Fink.
Special thanks to Gilad Schreiber for the web template.
project files:
- 'WEB files' - the web GUI and algorithms files.
- 'Scraping script' - simple scraping script used for collecting elements images (based on pinterest search queries)