This article focuses on a collection of early 20th century postcards from the Grand-Est region. These postcards feature, among other things, printed text giving details of a location, city or event. Many postcards also feature a postmark and postage stamps, which are of particular interest as they provide information on location and sending date. The aim of this article is to identify these postal markers to extract additional information, which can sometimes be missing from printed text descriptions. In order to detect, segment and recognize the contents of postal markers we combine two pre-trained neural networks models available in the literature. We report here the very promising first experimental results we obtained on a dataset of 1374 postcards.