ASSD: Attentive Single Shot Multibox Detector
This work addresses object detection for computer vision applications, but it is incremental as it builds on existing single-shot detectors with a novel attention mechanism.
The paper tackles object detection by proposing ASSD, a deep neural network that builds spatial feature relations to highlight useful regions and suppress irrelevant information, achieving competitive performance with state-of-the-art methods like SSD, DSSD, FSSD, and RetinaNet.
This paper proposes a new deep neural network for object detection. The proposed network, termed ASSD, builds feature relations in the spatial space of the feature map. With the global relation information, ASSD learns to highlight useful regions on the feature maps while suppressing the irrelevant information, thereby providing reliable guidance for object detection. Compared to methods that rely on complicated CNN layers to refine the feature maps, ASSD is simple in design and is computationally efficient. Experimental results show that ASSD competes favorably with the state-of-the-arts, including SSD, DSSD, FSSD and RetinaNet.