Image Tracking for the High Similarity Drug Tablets Based on Light Intensity Reflective Energy and Artificial Neural Network

Table 3

Representative recognition results of reflective energy surface properties for tracking those tablet templates on the 35th image frame.

Objective tablets	The recognized results (the obtained number vectors were the mean values based on 10 times of repeated experiments)		Recognition ratio
Template 1	0.00530.00240.00661.0035	Practical result vector	95.6%
	0001	Predetermined vector
Template 2	0.00140.00129.99710.0032	Practical result vector	94.6%
	0010	Predetermined vector
Template 3	0.0032−0.00111.00361.0042	Practical result vector	95.2%
	0011	Predetermined vector
Template 4	−0.00181.00510.00410.0017	Practical result vector	92.9%
	0100	Predetermined vector
Template 5	−0.00179.9981−0.00561.0088	Practical result vector	93.3%
	0101	Predetermined vector
Template 6	0.00551.00069.99680.0016	Practical result vector	95.1%
	0110	Predetermined vector
Template 7	−0.00551.00410.99881.0044	Practical result vector	98.2%
	0111	Predetermined vector
Template 8	1.00580.00660.00140.0058	Practical result vector	93.9%
	1000	Predetermined vector
Template 9	1.00550.00410.00880.9944	Practical result vector	95.6%
	1001	Predetermined vector
Template 10	1.00430.00220.9997−0.0035	Practical result vector	93.9%
	1010	Predetermined vector
Template 11	1.00130.00250.9927−0.0089	Practical result vector	94.4%
	1011	Predetermined vector
Template 12	1.00361.00220.0088−0.0071	Practical result vector	97.5%
	1100	Predetermined vector