A Convolutional Neural Network (CNN) has different implementations. AlexNet is one such implementation, and it won the ImageNet Challenge: ILSVRC 2012. Since then, CNNs have become omnipresent in computer vision and image detection and classification. Until April 2017, the general trend was to make deeper and more complicated networks to achieve higher accuracy. However, these deeper and complex networks offered improved accuracy but did not always make the networks more efficient, particularly in terms of size and speed. In many real-world applications, especially in IoT applications, such as a self-driving car and patient monitoring, recognition tasks need to be accomplished in a timely fashion on a resource-constrained (processing, memory) platform.
In this context, MobileNet V1 was introduced in April 2017. This version of Mobilenet...