• ACNN extracts spatial features, and strengthens relationships between channels. • ACNN equipped with characteristics of offline training and online learning performs outstandingly against state-of-the-art methods. • We have studied relations between the performance of the tracker and the number of layers. The moderate network achieves a good balance.
Attentional convolutional neural networks for object tracking
2018-04-01
2086428 byte
Aufsatz (Konferenz)
Elektronische Ressource
Englisch
Convolutional Neural Networks for Object Detection
SAE Technical Papers | 2024
|NEURAL NETWORKS WITH ATTENTIONAL BOTTLENECKS FOR TRAJECTORY PLANNING
Europäisches Patentamt | 2021
|Springer Verlag | 2017
|Spiking Deep Convolutional Neural Networks for Energy-Efficient Object Recognition
British Library Online Contents | 2015
|