×
About the project
Project Description
This project provides an HTML web page that allows users to visualize the output of a deployed computer vision DNN model. Users can improve on and gain insights from their deployed model by uploading query/test images and examining the model's results for correctness through the user interface. The web page includes some sample query/test images from the Microsoft image set, as well as example output for 3 types of models: Image Classification, Object Detection, and Image Similarity.
Usage
To use a deployed model in the Use My Model tab:
- Enter the model's API URL in the text field
- Upload or select images to use:
- Webcam:
- Allow the browser to use your web cam
- Select Snap Photo to take a picture
- Select Use Image to add the captured image
- Samples: Select an image by clicking on it
- Choose Files: Select images to upload from your machine's file explorer
- Select Upload to send the images to the model's API
- View results below!
To view examples in the See Example tab:
- Click on an image you wish to view
- See results from image classification, object detection, and image similarity models below!
Authors
This work was completed by a team of students from the Boston University College of Engineering as part of the EC528 Cloud Computing class. The project was completed in collaboration with three Microsoft engineers who proposed the project and acted as team mentors.
Student team:
Matthew Boyd, Charles Henneberger, Xushan "Mulla" Hu, SeungYeun "Kelly" Lee, Nuwapa "Prim" Promchotichai
Microsoft Mentors:
Patrick Buehler, Young Park, JS Tan