This is a repository for #AndroidDevChallenge
Generating pdfs from Videos work also need attention. The idea of this project is to convert video of a book pages to selectable pdf pages using on-device machine learning technique. Idea will work as follows:
- Start BookVideo2Pdf application and start video recording of a book for which you need to create pdf pages
- Hold page from a corner and try to turn, application will detect using on-device machine learning that this page should be converted to Pdf page
- PDF page will be created, which user can then select the text too.
All the pages will be appended, and in the end you will get the single pdf with out the need to capture images first.
All this work will help the users to generate selectable pdfs via capturing the videos of the books they want in a lesser time than first capturing.
Tell us what your idea is.
Start video recording of a book and turn the book page, application detects via on-device ML to convert this page to selectable text of pdf page. This application will help users in generating books to pdfs via videos in less time.
Tell us how you plan on bringing it to life. We believe that this idea has a lot of impact on the users. Currently, in order to create pdf of hard copies, users need to first capture the images, and then convert these images to PDFs, but limitation of this mechanism is that users would not be able to select the text available in the image. This idea has many advantages in terms of performance and effectiveness. Users will start video say at 30 fps, then, during video users will turn over pages, application will detect page turn and then text available in the frame will be converted to PDF as selectable text.
How machine learning is involved:
- If video is running at 30 fps and turning the page takes 1 second, then selecting a best frame
- Detecting the page turn during video
- Detecting page boundaries in the frame
Mustansar: Android Developer with 8 years of development experience
-
Survey conducting application https://play.google.com/store/apps/details?id=com.surveyauto.surveyauto:
-
Developed Pakistan's first School Education System ( https://play.google.com/store/apps/details?id=pk.gov.pitb.sis )
-
Developed Eye detection and recognition based attendance management system using IrisShield USB by Iritech, Inc. ( https://play.google.com/store/apps/details?id=pk.gov.pitb.saafpunjab.iris )
-
Developed Human Face detection based Camera using OpenCV4Android
-
Stackoverflow: https://stackoverflow.com/users/3027124/mustansar-saeed?tab=profile
-
Linkedin: https://www.linkedin.com/in/mustansar-saeed-3ba2941a/
Murtaza: Machine Learning developer with 3 years of experience
-
House detection and segmentation from satellite imagery
-
Image super-resolution
-
Dataset classification using neural network https://github.com/murtazakhan28/MNIST-dataset-classification-using-neural-network-in-python-and-numpy
-
Stackoverflow: https://stackoverflow.com/users/7621570/muhammad-murtaza-azam-khan
-
Linkedin: https://www.linkedin.com/in/murtaza-khan-63a844156/