SliceStream

SliceStream aims to develop an interactive user interface for high-quality segmentation and object tracking in videos, with a focus on reducing the reliance on labor-intensive data annotation and initialization. Our goal is to create a basic user interface linked with end-to-end machine learning models capable of targeting and tracking one object at a time.

Project Overview

Description

Current methods for video segmentation and object tracking require extensive manual annotation and initialization, limiting their scalability and practicality. SliceStream addresses this challenge by combining machine learning models with an intuitive user interface.

Expected Progress

Our primary goal for this semester is to create a basic user interface and integrate it with end-to-end machine learning models capable of segmenting and tracking a single object in videos.

Milestones

1. Literature Research and Model Selection (09/22/2023)

• Responsibility: Anil, Vincent, and Kevin
• Tasks: Research deep learning models, datasets, and metrics for object segmentation and tracking.

2. Front-End Design (09/29/2023)

• Responsibility: Vincent
• Tasks: Collaborate on the design and implementation of the front-end user interface.

3. Data Collection and Preprocessing (09/29/2023)

• Responsibility: Anil and Kevin
• Tasks: Collect and clean synthetic and real-world datasets for training and testing.

4. Segmentation Model (10/06/2023)

• Responsibility: Kevin
• Tasks: Select and validate state-of-the-art segmentation models on synthetic and real-world data.

5. Object Tracking Model (10/13/2023)

• Responsibility: Anil
• Tasks: Select and validate state-of-the-art object tracking models on synthetic and real-world data.

6. Integration (10/20/2023)

• Responsibility: Anil, Kevin, and Vincent
• Tasks: Combine segmentation and tracking models into an end-to-end solution. Train and validate on synthetic and real-world data.

7. Front-End UI (10/27/2023)

• Responsibility: Vincent
• Tasks: Design, test, and refine the user interface interaction.

8. Integration (11/10/2023)

• Responsibility: Anil, Kevin, and Vincent
• Tasks: Connect the front-end UI with the back-end models.

9. Documentation (11/17/2023)

• Responsibility: Anil, Kevin, and Vincent
• Tasks: Create technical documentation for the project.

10. Continuous Improvement (11/24/2023)

• Responsibility: Anil, Kevin, and Vincent
• Tasks: Gather user feedback and make necessary updates for ongoing improvement.

End of Semester (12/07/2023)

Stack

• Java
• Node.js
• Python

Repository

GitHub Repository

Team

• Anil Kamat
• Kevin Liu
• Vincent Lau

Meeting Space

• Pittsburgh 4206

Discord

Meeting Space on Discord