In this project, we were inspired by existing technologies such as Google Reverse Image search and facial recognition technologies. We wanted to incorporate the ability of computers to quickly identify objects in order to make searching for specific objects more easily.

What it does

Our project takes a picture of an object as an input. It analyzes the picture, and prints a short list of what the object most likely is in the picture. It then provides a link to Amazon search results based on the object that the picture could most likely be.

How we built it

To create this project, we found a model that was trained to identify objects. We used this model in conjunction with our code, so that our project would print out the most likely matches given by the model. We then used JavaScript to manipulate the web address to search Amazon for the object.

Challenges we ran into

We mostly struggled with our lack of experience with JavaScript and HTML/CSS. The majority of our programming experience comes from college courses, which were primarily taught in C++. Stepping away from C++ was something out of our comfort zone, but we were able to adapt to JavaScript and HTML/CSS to leverage their features.

Accomplishments that we're proud of

We are most proud of adapting to the two unfamiliar programming languages. Going into the project, we all had little knowledge of HTML/CSS, and even less so for JavaScript. Working on the project gave us a comfortable understanding of both languages after using them.

What we learned

This project taught how to manage projects on a larger scale than what we are all used to. Rather than try to solve single problems together, we divided tasks to accomplish the different parts of the lab simultaneously.

What's next for Image-based Amazon Search Engine

Our next step would be to refine the model for object recognition so that it would be able to identify objects more accurately.

Built With

Share this project: