Non-sequential realization methods

A part of the CVonline computer vision resource summarizing the different non-sequential algorithmic structures commonly encountered in computer vision and image processing.

1. Convolutional, Deep, Recurrent Neural Networks

   1. Convolutional neural network (CNN, or ConvNet)

   2. Deep neural networks

   3. Long short-term memory (LSTM)

   4. Recurrent neural network (RNN)

   5. Transfer learning

2. Systolic arrays

• 1. Digital signal processing/Digital signal processor

  2. Artificial neural networks

  3. Hopfield networks

  4. Kohonen networks

  5. Perceptron networks

  6. Radial basis function networks

  7. Support vector machines

3. Parallel processing approaches

   2. 1. GPU/GPGPU approaches

      2. Multiple Instruction Multiple Data (MIMD)/Multiprocessing

         1. Data parallelism/Domain decomposition methods

      3. Single Instruction Multiple Data (SIMD)

• • • 1. Massively parallel systems

3. 2. 4. Vector processors

4. Programmable logic approaches

• 1. Reconfigurable computing

5. VLSI approaches

• 1. Dynamic Vision Sensor

  2. Vision chips