Audiobook

While going on a walk or jogging I want to read a book however I can't do both at the same time from here a thought struck in mind what if I can listen to my PDF files as if someone is reading them to me? so that I can do both the activities (listening, walking) at the same time sounds exciting isn't

What it does

The Audiobook is an inspiration drawn from the above idea, the Audiobook can take a PDF file and outputs a human alike audio file you can choose the different voices based on your personal preference.

How we built it

I have built the software using various python modules. Initially, to process a page I used the pdf reader python module which parses the page and extracts the contents from the page. Once the page content is extracted if extracted content is text then I save it to a buffer file. However, sometimes a pdf file also contains images to extract the text from images I have used NLP API powered by Azure which extracts the general text and also handwritten text with accuracy greater than 85%

Challenges we ran into

The main challenge is to determine if the extracted content has image or text in-order to classify this I have used a text processor which only extracts text if the extracted content is null then I further parse the page to check if it has image. If an Image is detected, I use the pre-trained NLP model provided by Azure to extract the contents.

Accomplishments that we're proud of

I was able to successfully build software that takes a PDF file as input and provides the Audio file as output.

What we learned

I have learnt about text and image processing and how I can build it. Moreover, leveraging the pre-trained ML models to achieve the project goals

What's next for Audiobook

Audiobook will be coming up with a translator which converts the language as per the user preference.

Built With

• azure
• python