Skip to main content

Table 1 Comparison of two models mAP, recall, and mEAO

From: Efficient power component identification with long short-term memory and deep neural network

Model mAP mRecall mEAO
Mask R-CNN normal 0.793 0.82 0.83
Proposed model normal 0.833 0.85 0.91
Mask R-CNN disturbed 0.532 0.67 0.64
Proposed model disturbed 0.694 0.78 0.79