Inspiration
Similar images clubbed together for browsing always appeal the aesthetics rather than a random browsing order. After trying the app on movie names and famous wikipedia, we are trying to build something similar for Images.
What it does
Creates a meaningful browsing order for images based on similarity
How I built it
It uses Caffe-zoo neural net model to construct vector representation of images, K-nearest neighbours are calculated for any given image and a robust TSP tour is ran on the graph using a significant ranking. Imagenet data is used.
What's next for ImageBrowse
The next step would be to make it more personalized for each user similar to Google Photos
Built With
- caffe-model-zoo
- imagenet
- ipython-notebook
- python
Log in or sign up for Devpost to join the conversation.