This chapter focuses on the evolution of styles in the history of painting and retraces new genealogies of forms through recent deep learning-based computational methods applied to large image databases. This text associates the semiotic approach with the computational one offering a panorama of the most important current projects in Digital Art History. It proposes a particular conception of forms and forces in images, with the objective of making continuous bodily gestures analysable in large collections of paintings and photographs, starting from the works of art historians Henri Focillon and Aby Warburg. This conception of forces in images finds its roots in the Paris School Semiotics, but also in the work of philosopher Gilles Deleuze on pictorial diagrams in Francis Bacon’s painting and in mathematician René Thom’s idea of conflicting forces in painting.