Reading handwritten subscripts and superscripts in tesseract

[Sampurn Rattan Jain]

I want to read handwritten mathematical equations, which will not only contain regular the English alphabet, but also Greek  letter, along with numbers, subscript and superscript.

I have had some experience with reading handwritten notes using tesseract.

I have attached a sample image.

I am open to all advice, suggestions and guidance. 

[Jed Isom]

When I recently installed my version of tesseract, one of the language options was "Math / equation detection module".  Have you tried that yet?

[Sampurn Rattan Jain]

I am facing similar issues this guy is: https://groups.google.com/d/msg/tesseract-ocr/_V7pOll2kPo/JKkJGJMNqUAJ

That is, tesseract is returning garbage values.

I downloaded this training file from tesseract's git.

Ran the following command on this image.

tesseract -l equ ~/255286.png ~/one

cat ~/one.txt

∍⊥↙⊹≳−−∐ ∘↕ ∍⊥↙⊹≳− ∐

⇄ ⋮

⇄↸∂⊺≍≁⇄⊢≇≼−∐≱ ⇄↸∍⊺≍≁≇⊢⊈↻↥⊐

∍≍∙⊦⇂≨∶−∅≵ ∍≍∙⊦⇂≨∶⇄≀

∍∷≔ ≵∊ ∍∷≔↕⊠

≳∊ ∷∶∊

    

∷

My tesseract information:

tesseract 3.04.00

 leptonica-1.72

  libjpeg 8d (libjpeg-turbo 1.3.0) : libpng 1.2.50 : libtiff 4.0.3 : zlib 1.2.8