Automatic Recognition of Handwritten Bengali Broken Characters (BBC): Simulating Human Pattern Matching

摘要: This paper presents an automatic detection of handwritten Bengali Broken Characters (BBC) using a feed forward neural network (FFNN). It simulates the Human Visual System (HVS) way human eye matches patterns broken characters to meaningful character and identifies it. Here challenge is detect retrieve which has been distorted up 90%. The database consists fifty bangle characters, each with twenty samples. Each presented as image, preprocessed, segmented features are then extracted. A new method proposed in this paper. uses FFNN calculate mismatch for recognition character, where it observed that show very low original characters. For example, 70% distortions found be retrieved effectively.