From 4ee0b459256f641c3d179162a86c18cdea140691 Mon Sep 17 00:00:00 2001
From: summertight <soulvip1245@gmail.com>
Date: Sat, 21 Sep 2024 16:45:43 +0000
Subject: [PATCH] 123

---
 index.html | 21 ++++++++++++---------
 1 file changed, 12 insertions(+), 9 deletions(-)

diff --git a/index.html b/index.html
index 3eb29a6..f973a83 100644
--- a/index.html
+++ b/index.html
@@ -227,15 +227,18 @@ <h2 class="title is-3">Key Idea</h2>
         <a href="./static/images/fig_concept.pdf">Download PDF</a>.
       </iframe>
       <h2 class="subtitle has-text-centered">
-        (A) Conceptual comparison between prior works and our method. Prior works rely on a seesaw game of two potentially conflicting losses: reconstruction loss and identity loss. 
-        On the other hand, our method leverages a self-supervised approach with a clear ground truth, which allows for more stable training. 
-        
-        (B) Comparing our base approach (Ours Base) with our enhanced method (Ours Full), 
-        which includes techniques like perforation confusion and random mesh scaling. 
-        Green masks represent target-posed source 3DMM masks, red masks indicate target 3DMM masks, and orange masks denote their intersection.
-         The first row shows that when the source face is larger than the target’s, the jaw is cut off. The second row
-        shows the opposite case, where the base model fails to inpaint the remaining regions
-        effectively, while Ours Full generates realistic face-swapped outputs.
+        <p>
+          <strong>(A) Conceptual comparison</strong> between prior works and our method. Prior works rely on a seesaw game of two potentially conflicting losses: <strong>reconstruction loss</strong> and <strong>identity loss</strong>. On the other hand, our method leverages a <strong>self-supervised approach</strong> with a clear ground truth, which allows for more stable training.
+      </p>
+      
+      <p>
+          <strong>(B) Comparing our base approach</strong> (Ours Base) with our enhanced method (Ours Full), which includes techniques like <strong>perforation confusion</strong> and <strong>random mesh scaling</strong>. Green masks represent target-posed source 3DMM masks, red masks indicate target 3DMM masks, and orange masks denote their intersection.
+      </p>
+      
+      <p>
+          The first row shows that when the source face is larger than the target’s, the jaw is cut off. The second row shows the opposite case, where the base model fails to inpaint the remaining regions effectively, while <strong>Ours Full</strong> generates realistic face-swapped outputs.
+      </p>
+      
       </h2>
      
     </div>