fix link img

index.html

@@ -460,7 +460,7 @@ <h2 class="title is-3">Abstract</h2>
   </div>
   <div class="container is-widescreen" style="margin-top: 10px;">
     <div style="text-align: center;">
-      <img src="img/Teaser.jpg"  style="width: 100%;"">
+      <img src="./img/Teaser.jpg"  style="width: 100%;"">
     </div>
   </div>
 </section>
@@ -936,7 +936,7 @@ <h3 class="title is-4 has-text-centered">Setting Definition</h3>
             Furthermore, to concentrate on capturing the blur kernel, we create a known-blur images set (<img src="./img/mathK.png" width="16", height="16" style="margin-bottom: -2px;" alt="K">) by ultilizing the <i>Blur Kernel Extractor</i><sup>[2]</sup> - that can isolate and transfer blur kernels from randomly existing blurry-sharp image pairs to the targeted sharp inputs (<img src="./img/mathS.png" width="10", height="10" style="margin-bottom: -2px;" alt="S">).
           </p>
           <div style="text-align: center;">
-            <img id="method_definition" width="90%" src="./img/Overview.jpg" alt="Marigold training scheme"/>
+            <img id="method_definition" width="90%" src="img/Overview.jpg" alt="B2B_Overview"/>
           </div>
           <!-- <br> -->
 
@@ -946,7 +946,7 @@ <h3 class="title has-text-centered">Blur2Blur Pipeline</h3>
             This translator is trained using <span style="color: rgb(253, 127, 2);"><b>reconstruction</b></span> and <span style="color: purple;"><b>adversarial</b></span> losses. The converted images have known blur and can be successfully deblurred using the previously trained deblurring model.
           </p>
           <div style="text-align: center;">
-            <img id="method_pipeline" width="95%" height="auto" src="./img/B2B_Pipeline.jpg" alt="Marigold inference scheme"/>
+            <img id="method_pipeline" width="95%" height="auto" src="./img/B2B_Pipeline.jpg" alt="B2B_Pipeline"/>
           </div>
         </div>
       </div>