This chapter briefly introduces related basic concepts and algorithms and extracts key points for in-depth discussion. Section 2.1 introduces the framework of computer vision, focusing on a few key points for analysis, such as SIFT feature extraction, camera calibration, motion estimation and structure from motion (SFM), and other modules, such as stereo vision, shape from X, and object detection/recognition/tracking/segmentation, will be highlighted in Chap.  5 , so they are not repeated here; Sect. 2.1 briefly introduces image processing theory, especially analyzing two image denoising algorithms, namely bilateral filtering (BLF) and non-local means (NLM) filtering; Sect. 2.3 is an overview of optimization theory, especially discussing two common nonlinear least squares methods: G-N method and L-M method; Sect. 2.4 outlines the theory of machine learning, focusing on commonly used support vector machines and random forests; Sect. 2.5 provides an overview of deep learning theory, with a focus on convolutional neural networks (CNN), recurrent neural networks (RNN), generative adversarial networks (GAN), and Transformers; Sect. 2.6 gives the overview of neural network’s compression and acceleration techniques; Finally a introduction of efficient Transformers architecture applied in large-scale model theories and applications is given in Sect. 2.7.