diff --git a/README.md b/README.md
index c1a1a49..551dafe 100644
--- a/README.md
+++ b/README.md
@@ -61,7 +61,7 @@ After plotting the image, the automatic detection of ring borders can be perform
 t2 <- ring_detect(ring.data = t1, seg = 2, method = 'watershed')
 ```
 
-<center><img src="docs/articles/imgREADME-img001.png" width = "65%" height = "65%" /></center>
+<center><img src="docs/articles/img/README-img001.png" width = "65%" height = "65%" /></center>
 <center>Figure 1. The automatic detection of ring borders</center>
 
 ### 3. Calculate ring-width series 
@@ -90,7 +90,7 @@ This command allows to run a Shiny-based application within the system's default
 
 The dashboard has three components: a header, sidebar and body, like this
 
-<img src="docs/articles/imgREADME-img002.png" width = "85%" height = "85%" />
+<img src="docs/articles/img/README-img002.png" width = "85%" height = "85%" />
 
 A workflow for the Shiny app can be found at https://ropensci.github.io/MtreeRing/articles/app-MtreeRing.html. Most steps are demonstrated with a gif to make the workflow more understandable.
 
@@ -100,7 +100,7 @@ If an increment borer is used to extract samples, it is well known that the auge
 
 Under such conditions, you can create two paths by setting the argument `incline = TRUE`, or by ticking the checkbox **Inclined tree rings**. See this example.
 
-<img src="docs/articles/imgRingCorrection.png" width = "80%" height = "80%" /> 
+<img src="docs/articles/img/RingCorrection.png" width = "80%" height = "80%" /> 
 
 The line segment connecting two dots on the same ring should match the tangent of a tree ring border. The corrected ring width is estimated from the distance between adjacent rings and orientation of ring borders.