diff --git a/4-ComputerVision/12-Segmentation/SemanticSegmentationPytorch.ipynb b/4-ComputerVision/12-Segmentation/SemanticSegmentationPytorch.ipynb
index cc859e9e..2968bd9f 100644
--- a/4-ComputerVision/12-Segmentation/SemanticSegmentationPytorch.ipynb
+++ b/4-ComputerVision/12-Segmentation/SemanticSegmentationPytorch.ipynb
@@ -18,7 +18,7 @@
         "\n",
         "Almost all architectures have same structure. First part is **encoder** that extracts features from input image, second part is **decoder** that transforms this features into image with same height and width and some number of channels, may be equal to classes count.\n",
         "\n",
-        "<img src=\"images/segm.png\" width=\"50%\">\n",
+        "<img src=\"images/segm.png\" width=\"80%\">\n",
         "\n",
         "> Image from [this publication](https://arxiv.org/pdf/2001.05566.pdf)"
       ]
@@ -286,7 +286,7 @@
         "\n",
         "Simple encoder - decoder architecture with convolutions, poolings in encoder and convolutions, upsamplings in decoder.\n",
         "\n",
-        "<img src=\"images/segnet.png\" width=\"50%\">\n",
+        "<img src=\"images/segnet.png\" width=\"80%\">\n",
         "\n",
         "* Badrinarayanan, V., Kendall, A., & Cipolla, R. (2015). [SegNet: A deep convolutional\n",
         "encoder-decoder architecture for image segmentation](https://arxiv.org/pdf/1511.00561.pdf)"
@@ -552,7 +552,7 @@
         "\n",
         "U-Net usually has a default encoder for feature extraction, for example resnet50.\n",
         "\n",
-        "<img src=\"images/unet.png\" width=\"50%\">\n",
+        "<img src=\"images/unet.png\" width=\"70%\">\n",
         "\n",
         "* Ronneberger, Olaf, Philipp Fischer, and Thomas Brox. [U-Net: Convolutional networks for biomedical image segmentation.](https://arxiv.org/pdf/1505.04597.pdf)"
       ]