Precursors of CMEs in coronal images

For many years we have tried to find key observational signatures primarily in low coronal images (such as SOHO EIT) that are useful for predicting coronal mass ejections (CMEs). The signatures are isolated in an automated way. It is also important to estimate the spatial and kinematic properties of the white-light CME on the basis of its earlier signatures as captured in EUV and X-ray images. In this presentation, we give examples to demonstrate how low coronal observations help us predict or understand the spatial and kinematic properties of CMEs. We discuss how SDO AIA observations, either on their own or in combination with data from other experiments (such as STEREO), may improve our capability to predict CMEs and their properties.