A JavaScript library to find duplicate document images.

How does it work?

It extracts the text of images using OCR and uses levenshtein distance to calculate the similarity between two texts.

Online demo

• find. Find the duplicated images. You can pass your own OCR results.

  async find(images:HTMLImageElement[],textLinesOfImages?:TextLine[][],progressCallback?:any):Promise<HTMLImageElement[]>

• TextLine

  export interface TextLine{
    x:number;
    y:number;
    width:number;
    height:number;
    text:string;
  }

Via NPM:

npm install duplicate-document-image-finder

Via CDN:

<script type="module">
  import { DuplicateDocumentImageFinder } from 'https://cdn.jsdelivr.net/npm/duplicate-document-image-finder/dist/duplicate-document-image-finder.js';
</script>

MIT