diff --git a/index.html b/index.html
index 3c61191..459a337 100644
--- a/index.html
+++ b/index.html
@@ -85,7 +85,7 @@ <h4>Abstract</h4>
                 <h6 style="color:#8899a5" class="text-left"> 
                   We propose <strong>MASSTAR</strong>, a <strong>m</strong>ulti-modal l<strong>a</strong>rge-scale <strong>s</strong>cene dataset with a ver<strong>s</strong>atile <strong>t</strong>oolchain for surf<strong>a</strong>ce p<strong>r</strong>ediction and completion. 
                 </h6>
-                  <img src="assets/imgs/head_new.jpg" alt="Overview" width="70%">
+                  <img src="https://raw.githubusercontent.com/SYSU-STAR/MASSTAR/gh_page/assets/imgs/head_new.jpg" alt="Overview" width="70%">
                   <!-- <video width="100%" playsinline="" autoplay="autoplay" loop="loop" preload="" muted="">
                       <source src="video/teaser.mp4" type="video/mp4">
                   </video> -->
@@ -120,7 +120,7 @@ <h4>Multi-modal Dataset Generation Toolchain</h4>
                 <h5  class="text-center"> 
                   A. 3D Scene Segmentation
                 </h5>
-                <img src="assets/imgs/sam_new.jpg" alt="Overview" width="75%">
+                <img src="https://raw.githubusercontent.com/SYSU-STAR/MASSTAR/gh_page/assets/imgs/sam_new.jpg" alt="Overview" width="75%">
                 <br><br>
                 <p class="text-left">
                   An Overview of 3D Raw Data Segmentation. Initially, we generate the depth image and RGB image by rendering a bird's-eye view of the scene. Users have the option to employ SAM for segmenting top-view images through manual selection or fully automated methods. Subsequently, the 3D model is sliced using Blender Python, and CLIP is utilized to filter out non-architectural categories.
@@ -130,7 +130,7 @@ <h5  class="text-center">
                 <h5  class="text-center"> 
                   B. Images Rendering
                 </h5>
-                <img src="assets/imgs/render.jpg" alt="Overview" width="60%">
+                <img src="https://raw.githubusercontent.com/SYSU-STAR/MASSTAR/gh_page/assets/imgs/render.jpg" alt="Overview" width="60%">
                 <br><br>
                 <p class="text-left">
                   An example of the image rendering part of the toolchain. We offer the random mode (left) and trajectory mode (right) for users to choose from.
@@ -139,7 +139,7 @@ <h5  class="text-center">
                 <h5  class="text-center"> 
                   C. Descriptive Texts
                 </h5>
-                <img src="assets/imgs/text.jpg" alt="Overview" width="60%">
+                <img src="https://raw.githubusercontent.com/SYSU-STAR/MASSTAR/gh_page/assets/imgs/text.jpg" alt="Overview" width="60%">
                 <br><br>
                 <p class="text-left">
                   An example of the descriptive text render part of toolchain. we utilize BLIP to perform zero-shot image-to-text generation.
@@ -177,13 +177,13 @@ <h4>Example of dataset</h4>
                 <!-- <h6 style="color:#8899a5" class="text-left"> 
                   We propose MASSTAR, a multi-modal large-scale scene dataset with a versatile toolchain for surface prediction and completion.
                 </h6> -->
-                <img src="assets/imgs/rotation_new.gif" alt="ablation-representation" width="80%">
+                <img src="https://raw.githubusercontent.com/SYSU-STAR/MASSTAR/gh_page/assets/imgs/rotation_new.gif" alt="ablation-representation" width="80%">
                 <!-- <br> -->
                 <!-- <p class="text-center">
                    result 
                 </p> -->
                 <!-- <br> -->
-                <!-- <img src="assets/imgs/table.png" alt="ablation-representation" width="100%">
+                <!-- <img src="https://raw.githubusercontent.com/SYSU-STAR/MASSTAR/gh_page/assets/imgs/table.png" alt="ablation-representation" width="100%">
                 <p class="text-center">
                  the ouput .
                 </p>
@@ -208,7 +208,7 @@ <h4>Surface Completion Benchmark</h4>
                 <h6 style="color:#8899a5" class="text-left"> 
                   We conducted a comparative analysis of three surface prediction and completion algorithms, including SPM, PCN, and XMFnet.
                 </h6>
-                <img src="assets/imgs/result.gif" alt="ablation-representation" width="90%">
+                <img src="https://raw.githubusercontent.com/SYSU-STAR/MASSTAR/gh_page/assets/imgs/result.gif" alt="ablation-representation" width="90%">
                 <!-- <br>
                 <p class="text-center">
                    result 
@@ -217,20 +217,20 @@ <h6 style="color:#8899a5" class="text-left">
                 <!-- <div style="display: flex;">
                   <div style="display: flex; flex-direction: column;flex-direction: column;
                   justify-content: space-between;">
-                    <img src="assets/imgs/infer_time.png" alt="ablation-representation" width="90%">
+                    <img src="https://raw.githubusercontent.com/SYSU-STAR/MASSTAR/gh_page/assets/imgs/infer_time.png" alt="ablation-representation" width="90%">
                     <h8 class="text-center">
                       Time Efficiency and Resource Consumption of Each Method 
                     </h8>
                   </div>
                   <div style="display: flex; flex-direction: column;flex-direction: column;
                   justify-content: space-between;">
-                    <img src="assets/imgs/eval.png" alt="ablation-representation" width="90%">
+                    <img src="https://raw.githubusercontent.com/SYSU-STAR/MASSTAR/gh_page/assets/imgs/eval.png" alt="ablation-representation" width="90%">
                     <h7 class="text-center">
                       The Quality of Prediction Result of Each Method 
                     </h7>
                   </div>
                 </div> -->
-                <!-- <img src="assets/imgs/table.png" alt="ablation-representation" width="100%">
+                <!-- <img src="https://raw.githubusercontent.com/SYSU-STAR/MASSTAR/gh_page/assets/imgs/table.png" alt="ablation-representation" width="100%">
                 <p class="text-center">
                  the ouput .
                 </p>
@@ -252,7 +252,7 @@ <h2>Benchmark results</h2>
                 <hr style="margin-top:0px">
                 <video width="100%" playsinline="" autoplay="autoplay" loop="loop" preload="" muted="">
                     <source src="video/more_gso.mp4" type="video/mp4">
-                <img src="assets/imgs/result2.jpg" alt="ablation-representation" width="100%">
+                <img src="https://raw.githubusercontent.com/SYSU-STAR/MASSTAR/gh_page/assets/imgs/result2.jpg" alt="ablation-representation" width="100%">
                 </video>
                 <p class="text-center">
                   More results from GSO dataset.
@@ -277,14 +277,14 @@ <h2>Benchmark results</h2>
             <div class="col-12 text-center">
               <h2>Ablation</h2>
               <hr style="margin-top:0px">
-              <img src="assets/imgs/result2.jpg" alt="ablation-representation" width="100%">
+              <img src="https://raw.githubusercontent.com/SYSU-STAR/MASSTAR/gh_page/assets/imgs/result2.jpg" alt="ablation-representation" width="100%">
               <p class="text-center">
                 Qualitative comparison between different 3D representations: (1) naive generalizable 3D Gaussian (3DG), (2) Triplane-NeRF and (3) Triplane-Guassian.
               </p>
               <video width="100%" playsinline="" autoplay="autoplay" loop="loop" preload="" muted="">
                 <source src="video/ab.mp4" type="video/mp4">
               </video>
-              <img src="assets/imgs/plot_result2.jpg" alt="ablation-method" width="100%">
+              <img src="https://raw.githubusercontent.com/SYSU-STAR/MASSTAR/gh_page/assets/imgs/plot_result2.jpg" alt="ablation-method" width="100%">
               <p class="text-center">
                 Ablation experiments to assess the effect of Projection-aware Condition (P.C.) and Geometry-aware Encoding (G.E.).
               </p>