… that capsule network is … data may cause inaccurate reconstruction regularization values,
which may be a contributing factor to the subpar performance of capsule net on complex data. …

… -order data classification [20] such as 3-D HSI classification [21]. The strong capacity for
learning effective representations with multiple levels of abstraction for hyperspectral data allows …

… Even on the slightly more complex data set CIFAR10, the result of the capsule network is
only marginally comparable to, or even inferior to, the first generation of CNNs. Consequently, …

… classification of HSIs while significantly reducing the network design complexity. Specifically,
based on Hinton’s capsule … for HSI remote sensing data classification based on capsules […

… data, which is a common constraint in HSI [53]. There are some challenges for the existing
capsule-based networks for HSI classification… spatial data in a single classification framework (…

… Our DeeperCaps model successfully pushes the capsule network to a new level on Cifar-10,
a much bigger dataset, while also performing well on other datasets. By introducing the …

… This section details a brief introduction to capsule network architecture and its routing
mechanism and how our proposed method leverages existing work by taking advantage of this …

… with one category while encoding data in the capsule network. When compared to … capsule
network as independent combining of the encoded features is performed for every capsule. [6…

… of labeled data and an inefficient data routing policy, capsule networks … routing algorithm,
the capsule network has set the new … capsule network literature applied to image classification …

… , capsule networks not only learn good weights for feature extraction and image classification
… As an example, a capsule network learns to determine whether a plane is in the image, but …