Update instructions for using with KerasCV

README.md

@@ -210,20 +210,25 @@ ds = ds.padded_batch(batch_size=2, padding_values=-1.0)
 ```
 
 ### Using with KerasCV
-The [KerasCV](https://keras.io/keras_cv/) API has a little bit confusing format
-regarding the bounding boxes, depending on whether the users are preparing 
-the inputs for a preprocessing layer, or for a model.
+The [KerasCV](https://keras.io/keras_cv/) API is a little bit confusing 
+in terms of its input data format. The requirement is different between
+a preprocessing layer and a model.
 
-Targetran also provides easy conversion tools to make the process smoother.
+Targetran provides easy conversion tools to make the process smoother.
 ```python
 import keras_cv
 from targetran.tf import to_keras_cv_dict, to_keras_cv_model_input
 
-# The resulting dataset yields batches readily to be passed to a KerasCV
-# preprocessing layer. Batching in the appropriate format can be done directly 
-# with this call, therefore the above `padded_batch` example is not relevant here.
+# Let's assume `ds` contains Targetran ops as in the above illustration,
+# but without batching. To map the outputs to a KerasCV preprocessing layer,
+# the following can be done to get the required data format.
 ds = to_keras_cv_dict(ds, batch_size=2)
 
+# The resulting dataset yields batches readily to be passed to a KerasCV
+# preprocessing layer. Batching in the appropriate format is done directly 
+# with this call, therefore the `padded_batch` example is not relevant here.
+
+# Assume the user would like to add a jittered-resize op.
 jittered_resize = keras_cv.layers.JitteredResize(
     target_size=(640, 640),
     scale_factor=(0.8, 1.25),