Image is an important source of information for various applications, ranging from navigation, entertainment, through quality control in industry, to security systems. Exploitation of embedded imaging has been enabled by recent advances in technology of image sensors and embedded computing technology in both standalone and cloud-based systems. Machine learning plays a very important role in development of algorithms and applications in such systems. We will present some examples of exploitation of machine learning-based algorithms of image processing in both standalone embedded systems and cloud-based systems, such as object detection, HDR (High Dynamic Range) image acquisition and processing, and CNN (Convolutional Neural Networks) based image processing and improvement.