“…The translation of complex visual input into a phosphene percept (which by definition is limited) requires an efficient reduction of information and selection of the mere essential visual features for a given task. This can be achieved with the use of traditional computer vision approaches, such as edge detection ( Boyle, Maeder, & Boles, 2001 ; Dowling, Maeder, & Boles, 2004 ; Guo, Yang, & Gao, 2018 ), but deep neural network models have also gained increasing interest of prosthetic engineers (e.g., Sanchez-Garcia, Martinez-Cantin, & Guerrero, 2020 ; Han et al, 2021 ; Bollen et al, 2019 ; Bollen, van Wezel, van Gerven, & Güçlütürk, 2019 ; De Ruyter Van Steveninck, Güçlü, van Wezel, & Van Gerven, 2020 ; Lozano et al, 2020 ; Lozano et al, 2018 ). Various image processing approaches have been proposed for mobility in particular ( Barnes et al, 2011 ; Dagnelie et al, 2007 ; Dowling, Boles, & Maeder, 2006 ; Dowling, Maeder, & Boles, 2004 ; Feng & McCarthy, 2013 ; McCarthy et al, 2015 ; McCarthy, Feng, & Barnes, 2013 ; Parikh, Itti, Humayun, & Weiland, 2013 ; Srivastava, Troyk, & Dagnelie, 2009 ; van Rheede, Kennard, & Hicks, 2010 ; Vergnieux, Mace, & Jouffrais, 2014 ; Vergnieux, Macé, & Jouffrais, 2017 ; Zapf, Boon, Lovell, & Suaning, 2016 ).…”