Dynamic Multiscale Feature Fusion Method for Underwater Target Recognition

Table 3

Underwater blurred image recognition accuracy and recognition time.

Method	Torpedo	Torpedowake	Submarine	Frogman	Bubble	AUV	mAP	Time
FISHnet	0.5091	0.6616	0.8014	0.6402	0.6663	0.7525	0.6719	0.402
SiamFPN	0.3476	0.405	0.8459	0.6901	0.6645	0.8730	0.6377	0.228
SA-FPN	0.4134	0.4682	0.7786	0.7505	0.6517	0.8758	0.6564	0.254
Literature [33]	0.4588	0.4126	0.7075	0.7193	0.7204	0.8435	0.6436	0.104
Ours	0.4656	0.4189	0.8001	0.7759	0.7321	0.8306	0.6705	0.225

Note: the bold value in the table indicates excellent indicators of each algorithm.