Tactile Object recognition using low resolution image from close up image and Principle Component Analysis and z-score

Abstract

The development of an artificial sensory systems for robotic and computers is develop for a computer to recognize and understand the physical nature of the object. As The basis for the development of artificial intelligence system for computers. This paper presents a technique to recognize physical objects from the image tactile system by processing images from the object at close range. This study using a low resolution image size 16x16 point for 60 images per sample and for 6 groups with each group of 10 samples. Processing and analysis in this study, statistical features are extracted from a number of acquired tactile images for classification in their respective object image data. Principle Component Analysis (PCA) used for extract and analyzes an image sample.  Which resulted in average classification accuracy is 85%. The error will occur when the brightness of the object in the gray level that are very different as car sample is lowest accuracy is 50%.and ability grouping is accurate when the object is the brightness in the gray level are low differenced as tire sample is 100% accuracy.