Skip to main content

7-7 Instance Segmentation

PANDORA DEVELOPER GUIDE · CHAPTER 07

Learning Objectives​

Instance segmentation is a more fine-grained task in computer vision than object detection. It not only identifies what an object is and where it is, but also determines the precise shape of each object.

For a single image, the model outputs:

1. Object Category (What)​

For example: person, car, dog

2. Instance Mask​

Each object has a pixel-level binary mask

to accurately outlines the shape of the object

We will use Python and a pre-trained model to perform instance segmentation.​

Perform instance segmentation on an image and print the results.​

from ultralytics import YOLO
model = YOLO("yolo11n-seg.pt")
results = model.predict("https://ultralytics.com/images/bus.jpg")
for result in results:
    print(f'mask: {result.masks.data}') # Mask data (Number of objects x Height x Width)

The results are as follows: you will obtain a mask array with dimensions number of objects × height × width, where 0 indicates that the pixel does not belong to the object, and 1 indicates that the pixel belongs to the object.

7-7 Instance Segmentation

Perform instance segmentation using a webcam and visualize the results.​

import cv2
from ultralytics import YOLO
model = YOLO("yolo11n-seg.pt")
video_path = 0
cap = cv2.VideoCapture(video_path)
while cap.isOpened():
    success, frame = cap.read()
    if success:
        results = model.predict(frame)
        annotated_frame = results[0].plot()
        cv2.imshow("YOLO Inference", annotated_frame)
        if cv2.waitKey(1) & 0xFF == ord("q"):
            break
    else:
        break
cap.release()
cv2.destroyAllWindows()

The results are as follows.

7-7 Instance Segmentation

Adapted from the original YUAN Developer Hub article. Source ID: 261.

On this page