Image Master

Components:

---

1. Register Page:

The user has to enter basic details like username, email, password etc. The following Error cases are properly handled.

• User emails will be unique.
• Basic validation checks are handled.

2. Sign-In Page:

The user can sign in as either normal user role or admin role with the registered email address.

3. Home Page (Mode Selection Page):

The user after successfully logging in, can choose to join either a public or private channel to annotate images.

4. Classes Management Page:

• The admin can create new domains. He can create and add new classes under those domains.

• Admins can delete the domains or clases created by him/her.

• User can only annotate with an existing class that has already been created by the admin.

5. Public Dashboard:

• The list of all the images with their status (annotated or not) and annotator (the person who annotated the image) is displayed in this page.

• User can select a filter to view only annotated images or pending images.

• User can search for images by class name.

• User can upload new images into the public dashboard page. Multiple images can be uploaded at once (multiple formats are also supported).

• User can edit the status of annotation to the images that are previously annotated by him/her.
• User will not be able to edit the annotation of the images annotated by another user.
• The information of all the annotated images can be downloaded as a txt file.
• Users can select multiple images at a time to assign a class to all of them.

• Users can select multiple images at a time to assign a class to all of them (for both private and public dashboards).

5. Private Dashboard:

• The list of all the images that are owned by the user are displayed on this page. No user will be able to view the private dashboard of another user.

• User can select a filter to view only annotated images or pending images.

• User can upload new images into the public dashboard page. Multiple images can be uploaded at once.

Dockerized The Application:

Installation:

• 1. Go to Root directory after cloning the repository (ImageMaster)
• 2. Run command docker compose up --build in terminal
• 3. Server will run at localhost 5000, minio at 9001 and frontend at 5173
• 4. Then use the application at http://localhost:5173.

Without Docker Installation:

• 1. Clone the repository, install all dependency packages using npm install in both client and server folders.
• 2. Run minio server ./images to connect to minio client.
• 3. Run server using nodemon index.js and run client using npm start.

Dataset Used:

• For image functionality testing purpose, I used CIFAR 10 test images.

Demo Video:

• See the demo of Application here Click