Inspiration
We wanted to follow the StateFarm prompt to provide guidance for home maintenance. After discussing struggles with our own apartments, we ultimately decided on creating a web application to use AI to detect water damage.
What it does
The website is a stunning visual display that allows the user to upload a photo and have it analyzed by our machine learning algorithm.
How we built it
Initially, we struggled picking a platform. However, we settled on creating our own web server and hosting a website. We used Amazon AWS to create the server, then we installed Apache to host the website. We also used Flask to integrate Python into the web application.
Challenges we ran into
We could not get images to be shown by HTML and processed by Flask Flask is difficult to integrate in the web server Using CSS well is hard We had difficulty using content
Accomplishments that we're proud of
There are several accomplishes we're all happy to have accomplished. A few being integrating our trained AI imaging model, creating both a functional and proper website, and our choice to tackle the idea of detecting potential faults around customers houses.
What we learned
We've all had the amazing oppurtunity to utilize tools we haven't touched much before; from Flask to some of our team members learning HTML on the spot, it's all given us amazing insight into how we could all grow as hackers. Overall, we learned much about integrating AI models to HTML fueled website site through services and a complex yet beneficial python library.
What's next for Damage Detect Pro
We plan on doing a bit, we start slow to demonstrate the MVP to be benefical to customers, but we GREATLY intend to include multiple different things for our models to analyze. For example, we want to scan for foundation issues, shingle damage, and tile issues in relation to State Farm's insurance policies with respect to each category. We'd also like to be more explicit about our partnership with State Farm as we start implementing more tools and descriptions to our website. However, currently, we had neither the time, resources, nor people.
Log in or sign up for Devpost to join the conversation.