Inspiration
We want to support crowd sourced projects to improve software. All this can be done by exploring new paths and roads.
What it does
While being active outside and exploring new areas of our beautiful planet. One can contribute to solve issues in map features by image classification.
How we built it
We create an app that uses gamification to explore and take pictures of new areas while having fun and unlocking new levels. When the user takes a picture, it is send to the server where it can be classified as either a primary or a footpath. This information can then be used to update feature issues on open street map.
Challenges we ran into
As we were most unexperienced in app development we needed most of the time developing the app and connecting it to the server.
Accomplishments that we're proud of
We created a nice mockup where we implemented various gamification elements. Additionally, we not only trained a classifier on image recognition but also conducted hyperparameter-tuning and considered different models while only having limited compute resources and tight time schedule.
What we learned
A lot can be accomplished when working as a team and long dedicated hours.
What's next for Huawei
Implementing the mockup and using the predictions to solve issues for open street maps. Further hyperparamter-tuning can be done with more time and compute resources to improve the models performance.
Log in or sign up for Devpost to join the conversation.