NORTH ATLANTIC UNIVERSITY UNION
Automated wood recognition system is computationally expensive and require large data processing. The most time and resources consuming part of the system is classifier. Linear Discriminant Analysis (LDA) is used as classifier in existing wood recognition system on training stage and testing stage. LDA implementation to hardware is required for accelerating classifier performance. In hardware implementation, one of its important processes is floating-point to fixed-point conversion. Fixed-point number operations requires less computation complexities and hardware resources than floating-point. Fixed-point conversion is essentially process to determine optimum word-lengths of every involved variables.