Approximate Equivalents

Approximate Equivalents is a self-published 36-page book, printed in a limited series of 24. If you are interested in acquiring a copy, please email james[at]

These images are created by K.S.V. Judge, a machine learning model that selects and crops Google Street View panoramas of Pittsburgh to create photographic artwork. It was trained to emulate the compositional style and melancholic longing of Alfred Stieglitz’s Equivalents series. This book contains K.S.V. Judge’s favorite photographs, in order of increasing favor.

All images were upscaled using super-resolution machine learning models.

K.S.V. Judge was developed by James Thomas Wang, advised by David Oresick.