Centre for Eye Research Australia (CERA) - Research Publications

Permanent URI for this collection

Search Results

Now showing 1 - 1 of 1
  • Item
    Thumbnail Image
    Detection of available chairs with second generation suprachoroidal retinal prosthesis
    Moussallem, L ; Kolic, M ; Baglin, E ; Petoe, M ; Abbott1, C ; Lombardi, L ; Stefopoulos, S ; Battiwalla, X ; Walker, J ; Barnes, N ; Allen, P (Wiley, 2022)
    Purpose: Retinal prosthesis recipients have indicated that identification of chairs is a useful tool in the real-world environment. Current device software allows object localisation but does not specifically identify chairs. The aim was to compare accuracy in chair detection using a novel chair-specific vision processing method (ChD) to the current comprehensive vision processing method (Lanczos2; L2) in recipients of the second-generation (44-channel) suprachoroidal retinal prosthesis Method: Four implant recipients (#NCT05158049) with profound vision loss due to retinitis pigmentosa were acclimatised to both vision processing methods. Two mannequins (dressed in white or black) were seated face forward in two of three chairs (right, middle, left) in a square room (4 × 4m) with a white curtain backdrop. Participants (36 trials each, randomised) were asked to detect and navigate to the available chair (white or black) and navigation time was recorded. Results: The ChD vision processing method (87.5 ± 14.6% correct) performed significantly better than L2 vision processing method (19.4 ± 13.2%) for detecting available chairs (p = 0.020, Kruskal-Wallis). Furthermore, ChD performed better than L2 regardless of whether black (p = 0.019) or white (p = 0.015) chairs were used against the white backdrop. There was no difference in time taken to navigate to the available chair (L2 51.6 ± 25.1 s, ChD 48.4 ± 28.7 s; p = 0.564). Conclusion: The ChD processing method performs better than the L2 processing method for the purpose of specifically detecting chairs. Hence, there is potential for ChD to be incorporated into the bionic eye vision processing system to aid real-world navigation.