Add files via upload

index.html

@@ -232,10 +232,16 @@ <h2 class="title is-3">Abstract</h2>
             which are further incorporated with spatial features from the efficient point cloud processing network. 
             Feature aggregation is utilized to fuse them in a global-to-local strategy for compact segmentation features. 
             Experimental results show that our model outperforms baselines on both semantic and panoptic zero-shot segmentation task, 
-            meanwhile consumes less than 10% inference time compared to NeRF-based methods.
+            meanwhile consumes less than 10% inference time compared to NeRF-based methods, 
+            as shown in the below figure((b) and (c) are reported on Replica dataset).
           </p>
         </div>
       </div>
+      <p>
+        <img src="./static/images/demo.svg"
+        class="interpolation-image"
+        alt="Interpolate start reference image."/>
+      </p>
     </div>
     <!--/ Abstract. -->
 
@@ -261,6 +267,7 @@ <h2 class="title is-3">Method</h2>
         </figure> -->
       </p>
     </div>
+
   </div>
 </section>