Framework to Simulate Perceived Images Affected by Human Visual System Disorders

Human visual system (HVS) disorders, such as color vision deficiencies and impairments caused by traumatic brain injury (TBI), significantly alter perception. This paper presents a computational framework that leverages PerceptNet, a model of the HVS, combined with the optimization capabilities of deep learning frameworks such as JAX/Flax, to simulate how these disorders affect perceived images. The framework optimizes input images to match the feature maps of a modified PerceptNet model that incorporates disease-specific changes. We demonstrate its versatility by simulating visual perception in cases of color blindness and exploring the potential to model perceptual changes driven by neuroplasticity in the primary visual cortex (V1) following TBI. This framework offers a flexible and robust tool for studying altered visual perception and has potential applications in accessibility design, education, and research.Clinical relevance- This provides a framework to simulate how a disease in the human visual system changes the perception, providing both a system to adapt visual systems, and provides a computational framework to help study the behavior of some diseases.