Update README.md

README.md

@@ -24,7 +24,7 @@ size_categories:
 The dataset is oriented toward visual question answering of multilingual text scenes in nine languages, including Korean, Japanese, Italian, Russian, Deutsch, French, Thai, Arabic, and Vietnamese. The question-answer pairs are labeled by native annotators following a series of rules. A comprehensive description of the dataset can be found in the paper [MTVQA](https://arxiv.org/pdf/2405.11985).
 
 ## - Image Distribution
-<table style="width:50%;">
+<table style="width:60%;">
     <tr>
         <td><b>KO</b></td>
         <td><b>JA</b></td>
@@ -50,7 +50,7 @@ The dataset is oriented toward visual question answering of multilingual text sc
         <td>1139</td>
     </tr>
     <tr>
-        <td><b>Test images</b></td>
+        <td><b>Test Images</b></td>
         <td>250</td>
         <td>250</td>
         <td>250</td>