Scientific Reports

Table 2 Performance comparison of different methods on the dronevehicle validation dataset.

From: Cross-modal edge-enhanced detector for UAV-based multispectral object detection

DroneVehicle val.
Methods	Car	Truck	Freight-car	Bus	Van	mAP@50	Modality
RetinaNet⁴⁰	78.45	34.39	24.14	69.75	28.82	47.11
Faster R-CNN⁴¹	79.69	41.99	33.99	76.94	37.68	54.06
RoI Transformer⁴²	61.55	55.05	42.26	85.48	44.84	61.55	RGB
S2ANet²⁸	79.86	50.02	36.21	82.77	37.52	57.28
Oriented R-CNN²⁴	80.26	55.39	42.12	86.84	46.92	62.30
RetinaNet⁴⁰	88.81	35.43	39.47	76.45	32.12	54.45
Faster R-CNN⁴¹	89.68	40.95	43.10	86.32	41.21	60.27
RoI Transformer⁴²	89.64	50.98	53.42	88.86	44.47	65.47	IR
S2ANet²⁸	89.71	51.03	50.27	88.94	44.03	64.80
Oriented R-CNN²⁴	89.63	53.92	53.86	89.15	40.95	65.50
Halfway Fusion⁴⁵	89.85	60.34	55.51	88.97	46.28	68.19	RGB + IR
CIAN⁴⁶	89.98	62.47	60.22	88.90	49.59	70.23
AR-CNN⁴⁷	90.08	64.82	62.12	89.38	51.51	71.58
TSFADet⁷	89.88	67.87	63.74	89.81	53.99	73.06
Cacade-TSFADet⁷	90.01	69.15	65.45	89.70	55.19	73.90
CALNet⁴⁴	90.27	73.69	68.67	89.70	59.74	76.41
CMEE-Det	96.08	74.28	57.32	96.73	63.81	77.65

Back to article page

Search

Advanced search

Quick links