Back to Projects
Computer Vision

Image Segmentation with K-Means

Image segmentation using K-Means algorithm.

Course: Pattern Recognition in Data Mining
Image Segmentation with K-Means

Objectives

  • 1Segment images using K-Means algorithm.

Conclusions

  • Increasing K improves segmentation detail but introduces diminishing returns beyond K=8 for most images.
  • K-Means segmentation is computationally efficient, processing 512x512 images in under 1 second.
  • The algorithm struggles with objects of similar colors, as it only considers RGB color space information.

Technologies

  • OpenCV
  • FastAPI
  • NumPy
  • Matplotlib