🚀YOLOv8-Live-Instance-Segmentation-with-OpenCV

📌 Project Overview

This project implements a real-time instance segmentation system using YOLOv8 segmentation model and OpenCV. The system captures live video from a webcam, performs deep learning-based object segmentation, and displays segmented objects with pixel-level masks in real time.

🧠 Technologies Used

Python

YOLOv8 (Ultralytics)

OpenCV

Deep Learning (CNN-based detection & segmentation)

🔥 Features

Real-time webcam processing

Instance segmentation (pixel-level masks)

Multiple object detection

High-speed inference

Simple and clean implementation

🏗️ Project Architecture

Webcam → Frame Capture → YOLOv8 Segmentation → Mask Generation → Annotation → Display Output

⚙️ Installation

pip install ultralytics opencv-python

▶️ How to Run

python app.py

Press q to exit.

📊 How It Works

Capture frame using OpenCV

Pass frame to YOLOv8 segmentation model

Model predicts:

Bounding boxes

Class labels

Segmentation masks

Annotated output is displayed in real-time

🎯 Applications

Smart surveillance

Autonomous vehicles

Industrial automation

Retail analytics

Robotics vision systems

---

Project Architecture Explanation (For Interviews)

“This project is a real-time instance segmentation system built using YOLOv8 segmentation model. The pipeline starts with video capture using OpenCV. Each frame is passed to the YOLOv8 deep learning model. The model performs convolution-based feature extraction using a CNN backbone, then predicts bounding boxes, class probabilities, and pixel-level masks. The segmentation masks are then overlaid on the original frame, and the annotated output is displayed in real time.

The system runs in a continuous loop to achieve near real-time performance depending on hardware capabilities.”

---

👩‍💻 Author

Akshitha Hirakari