How do I use Python to generate images and strings for each product from linked menu images?

Sorry for the difficulty in understanding the title

From the menu image of Kura-zushi at the link below
http://www.kura-corpo.co.jp/menu-cp/img/menu-index-pc-01.jpg

·For each product (maguro|salmon|how much...)
·Character string (Taguro | 100 Yen | Offering area: All stores | Takeout: OK)
·Image (image of sushi for each product)

if you want to generate the ·Using what library
·How to recognize and cut out
Please let me know if it will be easy to implement.

When ordering sushi with your family,

Because it's troublesome to write down who asks for what and how many on paper. I wanted to create a system where I could easily count.

On a daily basis basis, running the Python script allows you to see
from the linked image. ·Image recognition(?)
·Writing characters(?)
·Perth, plastic surgery

to a table similar to the following ·INSERT, UPDATE
I'm imagining something to do.

+--------------------------------------------------------------------------------------------------------
| id | name | price | area | jpg_path |
+----+-------------+-------+------+------------------+
|  1 | maguro | 108 | all | hoge / maguro.jpg |
|  2 | salmon | 108 | all | hoge/salmon.jpg |
|  3 | ikura | 108 | all | hoge / ikura.jpg |
|  4 | new_sushiA | 108 | all | null |
...

Python scripts update tables daily
(If you have a new product, add a record and
I don't update the original product in particular) so
Based on that, generate html and
Tap the product to increase the count by one (Javascript?)
I would like to create a website like this

Because only Kurazushi is near my house

Usually, each product has a separate html element.
Use something like beautifulsup4 to parse and
I think I will collect strings and images, but
Kura-zushi wasn't like that, so I thought I'd try it.

Detecting people's faces with OpenCV,
There was a way to detect the outline, but

For OpenCV,
Recognize strings or
Cut out images of each product. ·Is it suitable for
·Are there any other good libraries?

For example,
The white background and the brown background are distinctly different, so
·Is there an easy way to recognize something other than that?
·If you cut it off with image editing software, it will be soon.

Forget about image recognition,
It's better to cut them into small pieces every certain length.

Should I start writing letters from the image?

I'd like to hear your opinion.
This is my first question, thank you.