How to parse recoIndex in enex files

According to this Evernote document, recoIndex-es in enex files store OCR scan result. However, they locate deep inside your notes in nested file structures. Actually, I wanted to make a GUI tool to access them easily but I could not spare them to prepare these interface and implement authentication procedures of Evernote. Then, the procedure, shown below, is the temporal answer to me to access recoIndex with ease of read.

1. Export your target note from any of Evernote client application.
2. Parse the exported enex file with the following command.

cat /tmp/YOUR_EXPORTED_ENEX_FILE_HERE | xmllint --xpath "//en-export/note/resource/recognition/text()" -|sed -e 's/^$//'|xmllint --pretty 1 -

Results are like this.

<?xml version="1.0" encoding="UTF-8" standalone="yes"?>
<!DOCTYPE recoIndex PUBLIC "SYSTEM" "http://xml.evernote.com/pub/recoIndex.dtd">
<recoIndex docType="unknown" objType="image" objID="a699335de622d35c5053785f405e639f" engineVersion="6.6.33.2" recoType="service" lang="ja" objWidth="2257" objHeight="2319">
  <item x="34" y="70" w="331" h="101">
    <t w="64">シナプス</t>
    <t w="54">INTス</t>
    <t w="53">ニンナ76ス</t>
    <t w="52">シナ76ス</t>
    <t w="52">シナク6ス</t>
    <t w="50">ニンナク6ス</t>
    <t w="21">16</t>
  </item>
  <item x="500" y="96" w="453" h="97">
    <t w="56">ニュータンの</t>
    <t w="54">ニュートンの</t>
    <t w="53">ニューロンの</t>
    <t w="53">ニューターの</t>
    <t w="52">ニュー畢ヽ一の</t>
    <t w="52">ニュータ一の</t>
    <t w="52">ニュー畢ヽーの</t>
    <t w="52">ニューpI一の</t>
    <t w="52">ニューpIーの</t>
    <t w="52">ニュー専ヽ一の</t>
  </item>