Machine Learning Model Morpheus Prepares to Analyze James Webb Telescope Images
Researchers are upgrading the Morpheus machine-learning model to classify galaxies in upcoming James Webb Space Telescope data. The software will help astronomers identify the oldest galaxies and study the evolution of dark matter.
Scientists are preparing to analyze the first images from the James Webb Space Telescope using an upgraded machine-learning model named Morpheus. This software automatically pores over telescope data to detect blurry blob-shaped objects in deep space and determines whether these structures are galaxies, classifying them by type.
The tool plays a crucial role in the COSMOS-Webb program, the largest project the telescope undertakes in its first year. A team of nearly 50 researchers relies on Morpheus to survey half-a-million galaxies as they hunt for the oldest fully-evolved galaxies to study how dark matter evolves over time.
Developers update Morpheus with modern attention methods that allow it to classify larger regions of images at once, speeding up the process by a factor of a hundred. The latest version also features new image processing capabilities like deblending, which separates overlapping astronomical objects to provide a clearer view of the deep universe.