Transfer Learning in Computer Vision: Technique and applications

Abstract

Transfer learning is a computer vision technique that allows a pre-trained model to be used as a starting point for solving a different but related task. It leverages the knowledge and learned representations from one task to improve the performance on another task. In practice, it is relatively rare to have a dataset of sufficient size to train an entire Convolutional Neural Network from scratch. Instead, it is common to pretrain a ConvNet on a very large dataset.