Intelligent Crack Detection and Quantification in the Concrete Bridge: A Deep Learning-Assisted Image Processing Approach

Figure 5

The performance of the network in the training process: (a) train box loss, (b) train objectness loss, (c) precision, (d) recall, (e) val box loss, (f) val objectness loss, (g) when the IoU threshold is 0.5, the calculated mAP (map@0.5), and (h) when the threshold of IoU ranges from 0.5 to 0.95 with a step of 0.05, the mean value of mAP (mAP@0.5 : 0.95).

(a)

(b)

(c)

(d)

(e)

(f)

(g)

(h)