128
правок
Изменения
Нет описания правки
:<tex>I_{comp}</tex> {{---}} изображение <tex>I_{gen}</tex>, в котором нестертые части заменены на части из <tex>I_{gt}</tex>; <tex>\Theta_{q}(x)</tex> {{---}} карта признаков, полученная <tex>q</tex>-ым слоем VGG-16.
* '''Style loss''' {{---}} сравнивает текстуру и цвета изображений, используя матрицу Грама<ref>[https://en.wikipedia.org/wiki/Gramian_matrix Gramian matrix, Wikipedia]</ref>.
<center><tex>L_{style} = \sum\limits_{q}\frac{1}{C_q C_q} \| \frac{G_q(I_{gen})-G_q(I_{gt})}{N_q}\|</tex></center>
== Примечания ==
<references/>
== Источники информации ==
* [https://heartbeat.fritz.ai/guide-to-image-inpainting-using-machine-learning-to-edit-and-correct-defects-in-photos-3c1b0e13bbd0 Guide to Image Inpainting: Using machine learning to edit and correct defects in photos, Heartbeat]
* [https://wandb.ai/site/articles/introduction-to-image-inpainting-with-deep-learning Introduction to image inpainting with deep learning, Weights & Biases]
* [https://towardsdatascience.com/pushing-the-limits-of-deep-image-inpainting-using-partial-convolutions-ed5520775ab4 Pushing the Limits of Deep Image Inpainting Using Partial Convolutions, Towards Data Science]
* [https://towardsdatascience.com/understanding-2d-dilated-convolution-operation-with-examples-in-numpy-and-tensorflow-with-d376b3972b25 Understanding 2D Dilated Convolution Operation with Examples in Numpy and Tensorflow with Interactive Code, Towards Data Science]
[[Категория: Машинное обучение]]
