Image registration

Image registration is a process whereby images are brought into spatial correspondence with one another. By registering images into a single coordinate system, one create fusion images and perform many quantitative analyzes.

Image registration is not restricted to aligning images of the same modality. Hybrid imaging requires image registration and is used to generate more informative images of pathology by combining complementary anatomical information across imaging modalities.

Most image registration algorithms consist of an iterative process that consists of three main steps: 1) transformation, 2) cost function and 3) optimization.

In image registration, the image we want to register is known as the native image (in its native coordinate space, usually referred to as native space) and the image we want to register to is known as the template image (in template space). To get an image from native space to template space, we need to apply a transformation. Examples of transformations include rotation, translation, shearing and scaling (or any combination of these). Image registrations restricted to rotation and translation are known as rigid body registrations, and registrations involving all four of these transformations are known as an affine registration.

Registration of images can also be described in terms of degrees of freedom, which can be given by the number of unique transformations multiplied by the dimensions of the image. For example, a 3D magnetic resonance image registered to a template using only rotation and translation is an example of a rigid body registration.

Once the transformation is applied, the quality of the registration is assessed using a cost function. The error in the iteration of the registration (quantified by the cost function) is then used to optimize transformation steps.

Related Radiopaedia articles

Artificial intelligence

• artificial intelligence (AI)
• imaging data sets
• computer-aided diagnosis (CAD)
• natural language processing
• machine learning (overview)
  • machine learning processes
    • supervised learning
    • unsupervised learning
      • clustering
    • reinforcement learning
    • evolutionary algorithms
    • validation split
  • machine learning models
    • linear regression
    • logistic regression
    • decision tree
    • random forest
    • support vector machine
    • feature scaling
    • linear discriminant analysis
    • neural network (overview)
      • neural network architectures
        • ​fully-connected neural network
        • recurrent neural network
        • convolutional neural network
• visualizing and understanding neural networks
  • generative adversarial networks (GANs)
  • heat maps
  • class activation mapping(CAM)
• common data preparation/preprocessing steps
  • DICOM to bitmap conversion
  • dimensionality reduction 
  • scaling
  • centering
  • normalization
  • principal component analysis
  • training, testing and validation datasets
  • augmentation
  • loss function
    • mean squared error
    • cross entropy
  • optimization algorithms 
    • ADAM 
    • momentum (Nesterov)
    • stochastic gradient descent
    • mini-batch gradient descent
  • regularisation
    • linear and quadratic
    • batch normalization 
  • ensembling
    • boosting
    • bagging
  • rule-based expert systems
• glossary
  • activation function
  • anomaly detection
  • automation bias
  • backpropagation
  • batch size
  • computer vision
  • concept drift
  • cost function
  • confusion matrix
  • convolution
  • cross validation
  • curse of dimensionality
  • dice similarity coefficient
  • dimensionality reduction
  • epoch
  • explainable artificial intelligence/XAI
  • feature extraction
  • federated learning
  • gradient descent
  • ground truth
  • hyperparameters
  • image registration
  • imputation
  • iteration
  • jaccard index
  • linear algebra
  • noise reduction
  • normalization
  • R (Programming language)
  • Python (Programming language)
  • segmentation 
  • semi-supervised learning
  • synthetic and augmented data
  • overfitting
  • transfer learning