Metabolomic data is conventionally analysed by assigning the peaks in the signal to different compounds, and this approach only takes into account a limited number of identifiable compounds. The aim of this study is to analysing metabolomic data by using a computer vision approach to take full advantage of deep learning and convolutional neural networks.