*4.1. Result*

Based on the CNN model, the loss of train data and test data was found to be 0.008 and 0.097, respectively, and the accuracy was 0.998 and 0.982, respectively (see Figure 5 and Table 9).

For more detailed learning results analysis, performance indicators by class were identified with train data and test data. First, we analyzed a confusion matrix for train data and checked the classification results by class. Confusion matrix is a matrix for comparing the predicted class with the actual class to measure the prediction performance through training, with the *x*-axis representing the predicted class and the *y*-axis representing the actual class. The results of the Confusion matrix (Figure 6) showed that some mutual confusion occurred between black ice and snow road, and when the actual class was wet road, it was predicted as road.

**Figure 5.** Training result: (**a**) The *x*-axis of the left graph represents the value of Epoch, the *y*-axis represents the value of loss; (**b**) the *x*-axis of the right graph represents Epoch, and the *y*-axis represents accuracy.


**Table 9.** Training Result.**Loss**

 **Accuracy**

> 0.998

> 0.982

 0.008

 0.097

**Class**

Train

Test

**Figure 6.** Confusion matrix; A matrix written to measure the prediction performance through training, the *x*-axis represents the predicted class and the *y*-axis represents the actual class. The results showed that (x,y) = (snow road, black ice) = 35, (black ice, snow road) = 11, (road, wet road) = 13.

Secondly, the calculation and analysis of accuracy, precision, and recall of each class was conducted on test data. The calculation results of each performance indicator are shown in Table 10. This shows that the accuracy of black ice, wet road and snow road is measured as relatively low, which is estimated to be the result of loss of light characteristics in the same way as the Confusion matrix analyzed earlier. However, the average values of accuracy, precision and call were 0.982, 0.983, 0.983, and 0.983, which are considered to have produced significant learning outcomes, even though the data are not relevant to learning.


**Table 10.** Accuracy, precision, recall results by class.
