diff --git a/chapters/en/4_Web_Of_Trust.md.txt b/chapters/en/4_Web_Of_Trust.md.txt
index eb0918de327e7802a894baa8144b2625866ff8e5..a5e00e99fe6111cfb0d0004265594577b3b827b3 100644
--- a/chapters/en/4_Web_Of_Trust.md.txt
+++ b/chapters/en/4_Web_Of_Trust.md.txt
@@ -142,7 +142,7 @@ alternatives being tested around the world for a decentralised communication net
 -   **In-degree of vertex A**: number of incoming edges / head ends
     to A.
 
-![degrees of a vertex diagram](./images/degres.jpg)
+![degrees of a vertex diagram](src/img/degrees_en.jpg)
 
 -   **Path**: -aka "walk"- path to follow to get from vertex A to
     vertex B.
@@ -347,7 +347,7 @@ well as that of the monetary community as a whole. The `xpercent`
 parameter prevents the creation of a 'faction' that could take hold of
 the blockchain.
 
-![Sybil region](./images/wot-sybil.jpg)
+![Sybil region](src/img/wot-sybil.jpg)
 
 The Sybil regions are isolated from the rest of the graph in the sense
 that they can only receive certifications from other ill-intentioned
@@ -386,7 +386,7 @@ Our goal with the Ğ1 is to create a community of about one million
 members to test the consequences of a libre monetary system. Let's see
 how we can tweak the pair of sigQty and stepMax- to reach this size:
 
-![Average WoT size graph as a function of sigQty and stepMax](./images/wot-moy.png)
+![Average WoT size graph as a function of sigQty and stepMax](src/img/wot-moy.png)
 
 The maximum size of a Sybil region grows linearly with `sigQty` but
 exponentially with `stepMax`. Logic has it that we need to keep
@@ -448,7 +448,7 @@ Here is a graph showing the evolution of a Sybil region with the
 variation of `sigPeriod`. The simulation considers that honest members 
 and attackers both issue a certification each `sigPeriod` interval, in days:
 
-![size of the WoT according to sigPeriod and stepAttackers](./images/impact_sig_period.png)
+![size of the WoT according to sigPeriod and stepAttackers](src/img/impact_sig_period.png)
 
 As we see, there is a strong link between the
 growth speed of the region and `sigPeriod`. As evidenced here, we need a
diff --git a/chapters/fr/4_Web_Of_Trust.md.txt b/chapters/fr/4_Web_Of_Trust.md.txt
index 622d1caccddc0f698b3c592ef9f730cb566c890c..515b4f439d7920ef88fa8bf6d502ec2c608738af 100644
--- a/chapters/fr/4_Web_Of_Trust.md.txt
+++ b/chapters/fr/4_Web_Of_Trust.md.txt
@@ -142,7 +142,7 @@ alternatives being tested around the world for a decentralised communication net
 -   **In-degree of vertex A**: number of incoming edges / head ends
     to A.
 
-![degrees of a vertex diagram](./images/degres.jpg)
+![degrees of a vertex diagram](src/img/degrees_fr.jpg)
 
 -   **Path**: -aka "walk"- path to follow to get from vertex A to
     vertex B.
@@ -347,7 +347,7 @@ well as that of the monetary community as a whole. The `xpercent`
 parameter prevents the creation of a 'faction' that could take hold of
 the blockchain.
 
-![Sybil region](./images/wot-sybil.jpg)
+![Sybil region](src/img/wot-sybil.jpg)
 
 The Sybil regions are isolated from the rest of the graph in the sense
 that they can only receive certifications from other ill-intentioned
@@ -386,7 +386,7 @@ Our goal with the Ğ1 is to create a community of about one million
 members to test the consequences of a libre monetary system. Let's see
 how we can tweak the pair of sigQty and stepMax- to reach this size:
 
-![Average WoT size graph as a function of sigQty and stepMax](./images/wot-moy.png)
+![Average WoT size graph as a function of sigQty and stepMax](src/img/wot-moy.png)
 
 The maximum size of a Sybil region grows linearly with `sigQty` but
 exponentially with `stepMax`. Logic has it that we need to keep
@@ -448,7 +448,7 @@ Here is a graph showing the evolution of a Sybil region with the
 variation of `sigPeriod`. The simulation considers that honest members
 and attackers both issue a certification each `sigPeriod` interval, in days:
 
-![size of the WoT according to sigPeriod and stepAttackers](./images/impact_sig_period.png)
+![size of the WoT according to sigPeriod and stepAttackers](src/img/impact_sig_period.png)
 
 As we see, there is a strong link between the
 growth speed of the region and `sigPeriod`. As evidenced here, we need a
diff --git a/publication.sh b/publication.sh
index 0e2a6a142ed29849d6870b338f3cc693da8a65c0..4fd34d8de5003560705885aaaabe8a0806830501 100644
--- a/publication.sh
+++ b/publication.sh
@@ -7,14 +7,11 @@
 
 # dépendances : pandoc
 hash pandoc 2>/dev/null || { echo >&2 "Ce générateur a besoin de la commande pandoc pour fonctionner. Mais cette commande n'est pas installée. Fin de l'exécution."; exit 1; }
-mkdir -p build/src/js
 ## Clean des dossiers de build et copie des assets
 [ -f whitepaper.html ] && rm whitepaper.html
 rm -rf build temp_whitepaper_html_en temp_whitepaper_html_fr temp_whitepaper_en.md temp_whitepaper_fr.md
-mkdir -p build/js
-#cp src/js/* build/js/
-mkdir build/img
-cp src/img/* build/img/
+mkdir -p build/src/img/
+cp src/img/* build/src/img/
 cp src/*.css build/
 
 ## Compilation des différents chapitres
diff --git a/src/img/degrees.jpg b/src/img/degrees_en.jpg
similarity index 100%
rename from src/img/degrees.jpg
rename to src/img/degrees_en.jpg
diff --git a/src/img/degrees_fr.jpg b/src/img/degrees_fr.jpg
new file mode 100644
index 0000000000000000000000000000000000000000..06024db146831c355480477fcd9f2224d88fae52
Binary files /dev/null and b/src/img/degrees_fr.jpg differ
diff --git a/src/img/impact_sig_period.png b/src/img/impact_sig_period.png
new file mode 100644
index 0000000000000000000000000000000000000000..d3f56fd62439344dc39d9599911e3837f5e14394
Binary files /dev/null and b/src/img/impact_sig_period.png differ
diff --git a/src/img/wot-moy.png b/src/img/wot-moy.png
new file mode 100644
index 0000000000000000000000000000000000000000..beaf76bb47226453cb18e1acddf7406661ee96d9
Binary files /dev/null and b/src/img/wot-moy.png differ
diff --git a/src/img/wot-sybil.jpg b/src/img/wot-sybil.jpg
new file mode 100644
index 0000000000000000000000000000000000000000..5c94b58b44db35387774b6bc823cf33bbe67a863
Binary files /dev/null and b/src/img/wot-sybil.jpg differ