The Research On Image-text Recognition And Reading Technology Based On Android Platform

According to statistics, there are around1.67%of the population in difficulty fornormal work and study due to visual or auditory disorders. The implementation ofimages-text recognition and reading technology reconstruct the eyes and ears of visual orauditory disorders. However, the existing image-text recognition and reading technology ismainly based on the personal computers, the integration of image-text recognition andreading technology based on mobile communication device (e.g.: Android phones) is not yetmature. Moreover, the text recognition process requires a big system overhead. With theemergence and development of the smart phone operating system, Android phones becomemore and more widely. How to implement image-text recognition and reading technologybased on Android mobile phone is a dream to all of visual or auditory disorders desiring, butalso a common problem the majority of researchers faced.Based on the aforementioned issues, in this paper, we implement a image-textrecognition and reading technology based on Android mobile platform that integrate Chineseimage-text recognition technology and English image-text recognition technology. Thus, thispaper implements Chinese and English image-text recognition function, and we improves thecharacter recognition rate by using an iterative method during character recognition. At thesame time, we also achieve a simultaneous reading function. The main processes for theimplementation of this technique: First, we pretreat the identified image, including: imagecropping, image gray processing and image binarization processing; Then, in order to savethe text recognition process overhead, this paper employs NDK mechanism andtesseract-two tools to recognize text in the image; Finally, this paper achieves asynchronization reading feature based on the android mobile client through speech synthesistechnology. And this feature has better scalability. We implement text synchronizationreading by calling the system provides an interface to use Pico engine, not only can beachieved on Chinese TTS speech synthesis can also be achieved in other languages. Inaddition, we also realize the image source selection function module, an image selected andpreview function module and display the results of conversion functional modules. Through testing each functional module of the system and performance of the core algorithm, theresults are as follows: image-text recognition and reading technology is faster, higher speedfor text recognition accuracy based on the Android platform achieved in this article, besides,the technology integrated better performance, with a certain significance and practical value.