Apr 20, 2020 · We propose the Orthogonal Softmax Layer (OSL), which makes the weight vectors in the classification layer remain orthogonal during both the training and test ...

The OSLNet uses the OSL as the classification layer instead of a fully connected classification layer, and the structure of the other layers are kept the same ...

Code release for OSLNet: Deep Small-Sample Classification with an Orthogonal Softmax Layer (TIP2020) - PRIS-CV/OSLNet.

The proposed Orthogonal Softmax Layer (OSL), which makes the weight vectors in the classification layer remain orthogonal during both the training and test ...

May 13, 2020 · Experimental results demonstrate that the proposed OSL has better performance than the methods used for comparison on four small-sample ...

A deep neural network of multiple nonlinear layers forms a large function space, which can easily lead to overfitting when it encounters small-sample data.

A deep neural network of multiple nonlinear layers forms a large function space, which can easily lead to overfitting when it encounters small-sample data.

A deep neural network of multiple nonlinear layers forms a large function space, which can easily lead to overfitting when it encounters small-sample data.

A deep neural network of multiple nonlinear layers forms a large function space, which can easily lead to overfitting when it encounters small-sample data.

Description : Deep Small-Sample Classification with an Orthogonal Softmax Layer. Code release for OSLNet: Deep Small-Sample Classification with an ...