rich pre-trained and out-of-box ai capabilities for ai newcomers
· handle challenging document situations such as rotation, occlusion, tilt, dense text, complex background, uneven lighting, dithering and blur.
· recognize special objects such as stamps, seals, handwriting, checkboxes, etc.
· support more than 10 common languages such as chinese (traditional and simplified), english, japanese, korean, french, etc.
· identify mixed layouts including framed, unframed tables, and text.
powerful customized ai capabilities and no code requirements allows easy creation of exclusive models according to business scenarios
· model-based and rule-based hybrid ai customization capabilities, using a small number of samples to customize the model.
· train the model using drag and drop operations to annotate element fields in documents such as invoices, receipts, orders, etc.
· classify both long text and short texts using only 20 pieces of training data per category.
· support a complete document processing workflow including data management, annotation, evaluation, launch functions while continuously optimizing the result through online training data.
long text comparison to accurately and rapidly process massive documents
· full-text content comparison of documents with different page numbers.
· support rapid positioning of differences in various locations through marking them in different colors.
· provide an intelligent intervention model to merge differences and remove redundant information such as spaces using rules of semantic information.
