diff --git a/launch/interaction/speech_recognition.launch b/launch/interaction/speech_recognition.launch
index 9e7b5d2..35d97d5 100644
--- a/launch/interaction/speech_recognition.launch
+++ b/launch/interaction/speech_recognition.launch
@@ -5,6 +5,7 @@
 	<arg name="machine" default="localhost"/>
     <arg name="dragonfly" default="true" />
     <arg name="kaldi" default="false" />
+    <arg name="qr_decoder_image_topic" default="~image" />
 
     <arg name="model_path" default="$(env HOME)/data/speech_models/model/"/>
 
@@ -18,6 +19,10 @@
 		<!-- QR-code / amigo-hear -->
 		<node name="string_topic_answerer" pkg="hmi" type="string_topic_answerer" output="log"/>
 
+		<node name="qr_code_decoder" pkg="hmi" type="qr_code_decoder" output="log">
+		    <remap from="~image" to="$(arg qr_decoder_image_topic)"/>
+		</node>
+
 		<!-- If we are on the real robot, launch the server, otherwise a dummy -->
 		<group if="$(optenv ROBOT_REAL false)">
 			<!-- bridge to windows -->