Introduction:
The study of the functional usages of urban greeneries are helpful in understanding their mental and physical health benefits on the community. Our project is developing an algorithm that classifies the function of an urban green, and predicts the activities taking place in it. Given pictures of the green space, our algorithm can predict multiple labels like [“art”,”kid”] if it is a good place for kids’ art events to happen.
Challenges & Insights: What has been the hardest part of the project you’ve encountered so far? Are there any concrete results you can show at this point?
We have been working on the preprocessing of data. The data we found online had multiple issues: file format (csv), duplicate images, different image sizes, incompatibility of images and labels, unreasonable label categorizations, etc. We had to program the fetching of images from the URLs, pad them to the same size, correspond the images and the labels, sort the labels into bigger and more general categories, and then save the images and labels to numpy arrays. This is now done, so our data is ready to be passed into a CNN model.
Plan: Are you on track with your project? What do you need to dedicate more time to? What are you thinking of changing, if anything?
We believe we are on track since we anticipated preprocessing taking up a big amount of time. Next, we are excited to implement our CNN model. Because of the structure of the data, our loss function will be pretty different from what we implemented in class: it will be a loss function that works with [input, [multiple labels]] pairs.
Log in or sign up for Devpost to join the conversation.