Abstract: Solutions for visual search and discovery include performing unsupervised training of a generative adversarial network that has a generator and an assessor. Training the generative adversarial network involves alternating training the assessor with the generator and a plurality of catalog images with training the generator with the assessor. The catalog images are inverted into catalog vectors by leveraging the trained generator. A query image is inverted into a query vector, and image similarity is determined by calculating a distance between the query vector and a catalog vector. In some examples, inversion is performed by training an encoder with the trained generator and inverting the catalog images with the encoder. In some examples, the trained generator is used to perform a search in a vector space. A weighting vector may be used to weight elements of the vectors, effectively prioritizing image features for image similarity determination.

Abstract: Systems and methods for identifying objects in an environment of a host vehicle are disclosed. In one implementation, a system includes a processor configured to receive images representative of the environment of the host vehicle; assign first pixel descriptor values to a plurality of pixels associated with a first image and second pixel descriptor values to a plurality of pixels associated with a second image; identify object representations in the first image and the second image based on at the first pixel descriptor values and the second pixel descriptor values, respectively; determine a first object descriptor and a second object descriptor based on the first pixel descriptor values and the second pixel descriptor values, respectively; and based on a comparison of the first object descriptor and the second object descriptor, output an indication that the object representations in the first image and the second image represent a common object.