Philosophical Poetry

• 

  Front page

• 

  Example poem

Inspiration:

Using cognitive services to create an unusual, unique and creative application that which uses the service in a way that is unexpected. We wanted to create a fun yet meaningful experience for the users. We wanted to marry literature and technology together which is not often explored.

What it does:

The application uses Microsoft Emotions API and Computer Vision API to explore emotions and features within any photograph to compose an interesting abstract poem.

How we built it

Back-end: Azure Emotions API to recognise emotions within portraits and people, Azure Computer Vision API to identify features within photographs. After identifying these words we used the word association API to compose relevant words related to the poems. Furthermore, we implemented our own poetic sentence structure to give the poems a natural feel and flow naturally.

Challenges we ran into

No one in our team had experience with Python so this was a huge learning experience. Furthermore, this was first time our team was using API's and integrating them together. Coming up with poetic sentence structures and associated words from the image key words proved to be the biggest challenge.

Accomplishments that I'm proud of

We manage to create cool poems most of the times! (Sometimes it can get a little weird though but that's ok, poems can be a little bit abstract :) )

What we learned

How effective and useful API's are to develop different functionalities of an application. Furthermore, this hackathon gave us a great idea of how Machine Vision works and how useful they can be in various applications. Additionally, we were grateful on the great documentation provided by Microsoft that made our life a lot easier using the API's.

What's next for Philosophical Poetry

We would like to create better sentence structure that is more accurate grammatically as well as making the poems more meaningful. Furthermore, we are thinking about machine learning techniques that could be used on a big data set of words to create predictions of which words are the most powerful in different types of poems.

Built With

• cognitive-services
• computer-vision-api
• django
• emotion-api
• python
• word-association-api