This article discusses the use of edge computing as a solution for processing large amounts of data from IoT devices, specifically focusing on image data. The proposed approach uses machine learning techniques for dimensionality reduction and has been shown to have minimal impact on accuracy.