Real-time image semantic segmentation network based on attention guidance mechanism

The invention discloses a real-time image semantic segmentation network based on an attention guiding mechanism. The real-time image semantic segmentation network comprises a down-sampling unit, an up-sampling unit, an extreme efficient residual module, a self-adaptive attention module and a self-adaptive fusion module. A feature extraction unit of the whole network structure is an extreme high-efficiency residual module, the calculation complexity of the module is effectively reduced by using an adaptive attention module (ASAM), and correlation information between effective pixel points can be captured; the low-level features and the high-level features are connected through an ASFM, and the features of different levels are connected in semantic segmentation; by stacking the five components, a real-time semantic segmentation network based on the attention mechanism is constructed, an encoder generates a down-sampled feature map, a decoder upsamples the deep feature map to match the resolution of an input ima