Optimize README

Signed-off-by: Jael Gu <mengjia.gu@zilliz.com>

README.md

@@ -76,13 +76,15 @@ It uses the pre-trained model specified by model name to generate an image embed
 
 **Parameters:**
 
-​	***img***: *numpy.ndarray*
+​	***img***: *[towhee.types.Image]()*
 
-​	The decoded image data in numpy.ndarray.
+​	The decoded image data in towhee.types.Image (numpy.ndarray).
 
 
 
-**Returns**: *numpy.ndarray*
+**Returns**:
+
+​   *numpy.ndarray*
 
 ​   The image embedding extracted by model.