A transformative and radically different approach for fracture characterization of soft adhesive materials through use of spiral cracking patterns is presented. This research particularly focuses on hydrocarbon polymeric materials such as asphalt binders. Five different asphalt materials were utilized in this study. An innovative integrated experimental-computational framework coupling Acoustic Emissions (AE) approach along with machine learning-based Digital Image Analysis (DIA) method is employed to determine the crack geometry and to accurately characterize fracture behavior of the material. Various image processing and machine learning techniques such as Convolutional Neural Networks (CNN), skeletonization, segmentation, and regression were used in DIA to automatically analyze spiral patterns. A new parameter called the “Spiral Cracking Energy (ESpiral)” to assess fracture performance of soft adhesives is introduced in this work. This study also explores the relationship between ESpiral and fracture energy of the material.