Question

The error matrix resulting from randomly selected test pixels for a classified image is given below. The Producer’s accuracy of class 1 is \rule{2.5cm{0.15mm} % (rounded off to 1 decimal place).} \begin{center} \begin{tabular}{|c|c|c|c|c|c|} \hline \multicolumn{2}{|c|}{} & \multicolumn{4}{c|}{Reference Data}
\cline{3-6} \multicolumn{2}{|c|}{} & Class 1 & Class 2 & Class 3 & Class 4
\hline \multirow{Classified Data} & Class 1 & 320 & 8 & 7 & 3
\cline{2-6} & Class 2 & 12 & 270 & 6 & 2
\cline{2-6} & Class 3 & 9 & 6 & 410 & 5
\cline{2-6} & Class 4 & 14 & 2 & 3 & 350
\hline \end{tabular} \end{center}

Hint:

Producer’s accuracy is computed by dividing the number of correctly classified pixels for a class by the total number of reference pixels for that class (i.e., column total).

Answer 1

Correct Answer: 90

Producer’s Accuracy is given by: \[ {Producer’s Accuracy} = \frac{{Correctly classified pixels in Class 1}}{{Total actual pixels of Class 1}} \times 100 \] From the matrix: - Correctly classified pixels in Class 1 = 320 (diagonal element) - Total actual pixels of Class 1 (column sum under Class 1) = \(320 + 12 + 9 + 14 = 355\) \[ {Producer’s Accuracy} = \frac{320}{355} \times 100 = 90.14% \] Final Answer: \fbox{90.1%}