journal menu
![[Figure 2]](yn5136fig2mag.jpg)
|
|
Figure 2
Schematic diagram of the GoniOwl CNN used for binary classification of sample-pin presence (pin_on versus pin_off) from RGB camera images (218 × 152 pixels). The network takes rescaled pixel values as input and comprises two convolutional stages, each containing two 3 × 3 convolutional layers with batch normalization and swish activation, followed by 2 × 2 max-pooling and dropout. The convolutional backbone is followed by global average pooling, two dense layers, and a sigmoid output neuron for binary classification. The architecture shown corresponds to the model selected in this study; the exact model definitions and training configuration are provided in the GitHub repository. |
![[Figure 2]](yn5136fig2.jpg)
 | JOURNAL OF SYNCHROTRON RADIATION |
ISSN: 1600-5775
access
