润色these methods often rely on the prior knowledge and low-level manual design features which limits the detection performance in complex scenes In addition the wealth of available information is ign

These methods often depend on prior knowledge and manual design of low-level features, which restricts their ability to detect objects in complex scenes. Moreover, they fail to utilize the abundant available information.