There are disclosed methods and apparatus for manufacture of image inventories from frames of a digital work including audio which corresponds to objects in still images of the digital video work. Objects are detected in each frame's image, and the objects are recognized. Metadata is assigned to the objects, the object metadata linking audio from the digital video work to the corresponding object in the frame's image which produces the audio. For each frame, at least one cryptographic hash of the object metadata is generated, and the hash is written to a node of a transaction processing network.
展开▼