← Back to Projects

Findr

Personal

September 2024 — December 2024

Overview

In my childhood, "I Spy" books have always been one of my favorite series. This fact played a large part into the decision for what my first large personal project would be. I wanted to make a program that would allow users to submit an image and a word. The software would then highlight the object indicated by user input.

Note, this project is incomplete, though, when developing, I have learned a lot about computer vision, deploying, and full-stack work, incorporating many new technologies. The reason for failure was due to no access to a large dataset of "I Spy" content for training, as well as no motivation to create the dataset for myself.

Findr Demo

Key Details

  • Frontend Development: Having done mostly backend work prior to this project, I gained valuable skill in frontend development. In researching modern frontend frameworks, I decided to use Angular. I also read up on JAMstack architecture as well as associated documentation.
  • Computer Vision: The core focus of this project was on computer vision and learning how exactly this is accomplished. I learned about datasets and big data, as well as public dataset hubs like Kaggle and Hugging Face.
  • APIs: With no prior API experience or knowledge, I researched into RESTful APIs and backend frameworks. I stumbled upon flask and Django REST framework.
  • Software Deployment: Again, with no prior deployment experience or knowledge, I researched into AWS for both deployment of the app, but also Sagemaker for the potential deployment of the machine learning model.