What's new

Identifying Kanji Online With Stroke Recognition

Today I want to write about identifying Kanji online with stroke recognition. Although it takes longer than using furigana, it can be very helpful when all you have are the actual kanji.

While it can be easy to determine that the first line indicates: "November 22nd, 8:43 in the morning (午前)", the second part contains a lot more unfamiliar kanji.
I'll break down the second line in three portions (each segment separated by a space).
Stroke Recognition:
So in the first segment, I'm looking for the two remaining kanji. I could use stroke recognition in order to deal with the second one. All I need to do is to write the kanji I'm looking for in the box and then select the kanji I'm looking for like so:

Once I have access to my kanji, I can simply copy/paste it to me text in order to do my research:
This is rather time consuming however and writing with a mouse can be difficult.
Stroke Number Lookup:
Another option is to lookup a kanji by the amount of strokes required to write it. So the third kanji I'm looking for is:

The "B" shaped radical on the left is three strokes. The right part is seven strokes. This makes a total of 10 strokes. On Jim Breen's website I can simply fill in the search info (1 & 2) and then lookup the kanjis (3) available to make my selection:

Unfortunately, there are a total of more than 500 kanjis which meet that criteria which makes it even more time consuming to find.

Radical Lookup:
By using dictionary software, I can use the radicals found in the kanji to quickly narrow my search:


Having now access to all three kanji:
大審院, I can finally find out the meaning of the word:

Radical Number Lookup:
Kanji radicals have a number assigned to them:

Going back to our original script, the first kanji of the second segment is:

I've identified the radical on the left part of the character here. We can then lookup that radical number:

By putting "145" in my radical number search box, I'll quickly have access to a lot less potential kanji:


大審院 被...
Using OCR to Recognize Kanji
Moving on to the next character in our script:

I've tried various online websites which boast being able to recognize pictures from image files so I submitted the files above. Of the 10 websites or so which I tried, I couldn't find a single one which worked. If you know of an OCR which works, let me know and I'll try it out.
大審院 被告...
As the next one down the line is fairly easy to identify, I'll add it in as well: 人.
大審院 被告人

Looking up a Kanji with Multi-Radicals
I've already covered this way of proceeding in a previous entry but here it is again if you'd like to do it online.
By breaking down a single kanji by its constituents, it's then possible to tick them from a list and then get the kanji you're looking for:


In order to get:

大審院 被告人...
Using a Header and a Book
Identified below is the next kanji to be identified. In the book The Key to Kanji, kanjis are organized along various components including "headers". The headers here being the upper part of the kanji:

Following the index, it's then fairly easy to identify the kanji by looking up the reference in the book:
室:シツ⋆むろ: a room.
The book provides both on-yomi and kun-yomi as well as an explanation about the etymology of the kanji. Unfortunately, it "only" covers 1100 kanji.
大審院 被告人控室...
As for the remaining kanji, one has the furigana provided: 伍号室
大審院 被告人控室 伍号室:
Supreme Court Defendant Waiting Room #5.
I hope this helps and thanks for reading!


There are no comments to display.

Journal entry information

Last update

More entries in Blogroll

More entries from Zizka