Update README.md

README.md

@@ -18,12 +18,16 @@ features. We define two tasks, which are to preserve gaze information and to rem
 information. The two tasks are not cooperative but adversarial to purify feature. Simultaneously optimizing the two tasks, we
 implicitly purify the gaze feature without defining gaze-irrelevant feature.
 
-<div align=center>  <img src="figures/pipeline.png" alt="performance" width="800" align="bottom" /> </div>
+<div align=center>  <img src="figures/performance.png" alt="performance" width="800" align="bottom" /> </div>
 
 **Performance:** PureGaze shows best performance among typical gaze estimation methods (w/o adaption), and has competitive result among domain adaption methods. Note that, PureGaze learns one optimal model for four tasks, while domain adaption methods need to learn a total of four models. This is an advantage of PureGaze.
 
-<div align=center>  <img src="figures/visual.png" alt="performance" width="800" align="bottom" /> </div>
+<div align=center>  <img src="figures/visual.png" alt="visualization" width="800" align="bottom" /> </div>
 
 **Feature visualization:** The result clearly explains the purification. Our purified feature contains less gaze-irrelevant feature and naturally improves the cross-domain performance.
 
-## Resource
+## Requirement
+We build the project with   `pytorch1.7.0`.
+
+
+