I've been trying to understand how to configure tesseract options for page segmentation. I attempted to use tessedit_pageseg_mode: '1', but encountered a halt in the text recognition process. If I input it as number 1, the process completes, but the mode remains default SINGLE_BLOCK.
Currently, I'm running Tesseract version: 1.0.19. I also experimented with 2.0.0-alpha.13, with identical results.
this.tesseract = Tesseract.create({
workerPath: '../../assets/tesseract/worker.js',
langPath: '../../assets/tesseract/trained-data',
corePath: '../../assets/tesseract/core.js',
});
this.tesseract.recognize(this.image, {
lang: 'eng',
tessedit_pageseg_mode: '1'
})
.progress((p) => {
console.log('progress', p);
this.ocrResult = p.status + ', Progress: ' + Math.round(p.progress * 100) + '%';
})
.then((data) => {
console.log(data.psm);
this.ocrResult = data.text;
}).catch((err) => {
console.error('Error occurred while recognizing text', err);
});
Your assistance would be highly appreciated. Thank you!
Update: I managed to identify the problem. My previous code was written like this
window.Tesseract = Tesseract.create({...Tesseract.recognize(.. which led to downloading the worker, language, and core files from the internet instead of utilizing the provided file path. The file download was somehow getting interrupted when I included the option tessedit_pageseg_mode: '1'.
Uncaught abort() at Error at Na (https://cdn.jsdelivr.net/gh/naptha/[email protected]/index.js:36:26)
Therefore, I am now attempting to utilize local files by adding file:///android_asset/www (as described in this blog post) within the assets folder. However, I am encountering an issue where it fails with this message.
Failed to load file:///android_asset/www/assets/plugins/tesseract/worker.js: Cross origin requests are only supported for protocol schemes: http, data, chrome, https.`