Segmentation Techniques

Duration: 7 min

This module delves into segmentation techniques, a crucial aspect of computer vision that involves partitioning an image into meaningful segments. Understanding segmentation is vital for applications like medical imaging, autonomous driving, and image editing, where precise delineation of objects is necessary.

Understanding Image Segmentation

Image segmentation is the process of dividing an image into multiple segments or regions, each representing a different object or area of interest. This technique is essential for detailed image analysis and is used in various applications, including object recognition, scene understanding, and medical imaging.

import cv2
import numpy as np

# Load an image
image = cv2.imread('example.jpg')

# Convert to grayscale
gray = cv2.cvtColor(image, cv2.COLOR_BGR2GRAY)

# Apply thresholding
_, thresholded = cv2.threshold(gray, 127, 255, cv2.THRESH_BINARY)

# Find contours
contours, _ = cv2.findContours(thresholded, cv2.RETR_EXTERNAL, cv2.CHAIN_APPROX_SIMPLE)

# Draw contours on the original image
cv2.drawContours(image, contours, -1, (0, 255, 0), 2)

# Display the image with contours
cv2.imshow('Segmented Image', image)
cv2.waitKey(0)
cv2.destroyAllWindows()

Try it in Google Colab:

Displays an image with contours drawn around segmented regions.

Deep Learning-based Segmentation with U-Net

U-Net is a popular convolutional neural network architecture for image segmentation, especially in medical imaging. It consists of a contracting path to capture context and a symmetric expanding path that enables precise localization. U-Net is known for its ability to produce high-quality segmentation results with relatively few training images.

import tensorflow as tf
from tensorflow.keras.models import Model
from tensorflow.keras.layers import Input, Conv2D, MaxPooling2D, UpSampling2D, concatenate

# Define the U-Net model
def unet(input_size=(128, 128, 1)):
    inputs = Input(input_size)
    # Contracting path
    c1 = Conv2D(64, (3, 3), activation='relu', padding='same')(inputs)
    c1 = Conv2D(64, (3, 3), activation='relu', padding='same')(c1)
    p1 = MaxPooling2D((2, 2))(c1)
    
    # Expanding path
    u1 = UpSampling2D((2, 2))(p1)
    u1 = concatenate([u1, c1], axis=3)
    c2 = Conv2D(64, (3, 3), activation='relu', padding='same')(u1)
    c2 = Conv2D(64, (3, 3), activation='relu', padding='same')(c2)
    outputs = Conv2D(1, (1, 1), activation='sigmoid')(c2)
    
    model = Model(inputs=[inputs], outputs=[outputs])
    return model

# Create and compile the U-Net model
unet_model = unet()
unet_model.compile(optimizer='adam', loss='binary_crossentropy')

💡 Tip: When training U-Net, ensure that your dataset is properly preprocessed and augmented to avoid overfitting, especially when working with limited data.