Author: Kelvin Luu

This is a continuation of this post [link!] detailing our investigation of thresholding as a means to silhouette our images. In this post, we will see if using a contemporary machine learning model for image segmentation will yield better results. Introduction With the thresholding being prone to fairly disappointing results, a colleague recommended that we look into Meta/Facebook’s machine learning Segment Anything model, which is thankfully open source. We ideally did not want to rely on a program that could be taken down or otherwise become unusable. We want to be able to do this long-term, so this being open gave us more confidence in its sustainability. Setup The README for the repo is fairly detailed about getting dependencies and the actual model installed. The steps are fairly standard aside…

Introduction One of the questions we get periodically in the Digital Archaeology Lab is around photographing artifacts and silhouetting the objects (aka removing the background so that you have an image of just the object that can then be used in publications or presentations). In most cases, this process has been done manually using software like Adobe Photoshop or GIMP to digitally draw a mask over the object and then deleting the background pixels. However, this does not scale and there has always been this nagging thought of shouldn’t technology be able to do this for us by now? The Carpentries team published a workshop on image processing that reignited this thought and showed some possibilities for ecological data. Yet we know that the range of materials, composition, conditions, and…