HomeView

Reconstruction
triangulation
homeview logo

Inspiration

As a person who likes traveling and has been to multiple different countries, one of the biggest issues in traveling has been the housing. Home images on listings often seem bigger or brighter in pictures, but the reality is often disappointing. Our inspiration comes from real life experience, which is why we wanted to make an website that allows travelers, renters, and homebuyers the ability to explore properties in a more accurate and immersive way, ensuring that what they see is truly what they get.

What it does

Homeview is a platform that allows users to explore real estate properties through immersive 3D models. By offering accurate, detailed models, HomeView ensures that buyers, renters, and travelers can view homes as if they were physically there. This helps users make more informed decisions, eliminating the discrepancies between online images and the actual property.

How we built it

The app consists of only two parts. The first part consists of the front end written in Nextjs. It uses MangoDB Atlas and Edgestore for storage and Github0 Auth for authentication. It is intened to be the primay form of connecting to our app.

The backend is written primarily in python, but uses a lot of code transpiled from c++. This includes tools such as opencv and pandas for image manipulation and data management, but also SFM libraries such as colmap.

Challenges we ran into

We ran into challenges such as :

Performance issues: compiling and running the program, often resulted in structures taking too long to generate.
Complex coding and debugging: Implementing dynamic elements can often lead to unexpected bugs or crashes.
Compatibility across platforms: working across multiple devices and platforms is challenging
Resource constraints: time and budget made it hard to execute the desired effects.