Image Marker (2507.02153v1)

Abstract: A wide range of scientific imaging datasets benefit from human inspection for purposes ranging from prosaic-such as fault identification and quality inspection-to profound, enabling the discovery of new phenomena. As such, these datasets come in a wide variety of forms, with diverse inspection needs. In this paper we present a software package, Image Marker, designed to help facilitate human categorization of images. The software allows for quick seeking through images and enables flexible marking and logging of up to 9 different classes of features and their locations in files of FITS, TIFF, PNG, and JPEG format. Additional tools are provided to add text-based comments to the marking logs and for displaying external mark datasets on images during the classification process. As our primary use case will be the identification of features in astronomical survey data, Image Marker will also utilize standard world coordinate system (WCS) headers embedded in FITS headers and TIFF metadata when available. The lightweight software, based on the Qt Framework to build the GUI application, enables efficient marking of thousands of images on personal-scale computers. We provide Image Marker as a Python package, and as Mac and Windows 11 executables. It is available on GitHub or via pip installation.