Abstract

Info

Title:: Image processing strategy for object recognition in artificial vision based on salient object detection

Author(s):: ZHANG Yan1; ZHAO Ying2; CAO Feng3; JIANG Guangmiao2; HE Yang2; WANG Sheng2; WANG Nan1; 1. School of Computer and Information Engineering, Qilu Institute of Technology, Ji’nan 250200, China; 2. School of Digital andIntelligent Industry, Inner Mongolia University of Science and Technology, Baotou 014010, China; 3. School of IntelligentManufacturing and Control Engineering, Qilu Institute of Technology, Ji’nan 250200, China

Keywords:: visual prosthesis; simulated prosthetic vision; salient object detection; object recognition; image processing strategy

Abstract:: Abstract: Objective To propose a image processing strategy based on salient object detection algorithm for optimizing thepresentation of prosthetic visual information at a limited resolution level, aiming to detect and enhance the salient objects inthe scene and remove the background information. Methods A salient object detection model combining CNN andTransformer was used to extract salient objects. On this basis, methods such as object magnification, contour enhancementand contrast enhancement were utilized to optimize the image information. Psychophysical experiments were carried out at 5resolution levels (16×16, 24×24, 32×32, 48×48 and 64×64). Results In the simulated prosthetic vision, this image processingstrategy had a remarkable effect on improving the object recognition ability of the subjects. Regardless of the resolutions of16×16, 24×24, 32×32, 48×48 and 64×64, the proposed strategy achieved the highest recognition accuracies, specifically34%±6%, 56%±9%, 72%±9%, 89%±4% and 96%±2%. Conclusion Using the salient object detection method and imageprocessing strategy to extract and enhance salient objects can help prosthesis implant recipients effectively improve theirobject recognition ability.