{"id":11150,"date":"2026-08-18T13:54:51","date_gmt":"2026-08-18T05:54:51","guid":{"rendered":"https:\/\/infernews.com\/blog\/navidc-ocr\/"},"modified":"2026-08-18T13:56:18","modified_gmt":"2026-08-18T05:56:18","slug":"navidc-ocr","status":"publish","type":"post","link":"https:\/\/infernews.com\/blog\/navidc-ocr\/","title":{"rendered":"NaviDC-OCR\uff1a1.2B \u53c3\u6578\u6587\u6a94 VLM\uff0c\u7d71\u4e00\u8655\u7406\u76f8\u6a5f\u8207\u6578\u78bc\u6587\u4ef6"},"content":{"rendered":"\n<figure class=\"wp-block-image size-large\"><img decoding=\"async\" src=\"https:\/\/infernews.com\/blog\/wp-content\/uploads\/2026\/08\/icon.jpg\" alt=\"icon\"\/><\/figure>\n\n\n\n<p class=\"wp-block-paragraph\">NaviDC-OCR \u662f\u4e00\u500b\u8f15\u91cf\u7d1a Vision-Language Model\uff08VLM\uff09\uff0c\u5c08\u9580\u8655\u7406 document understanding\uff0c\u4e5f\u5c31\u662f\u628a\u6587\u4ef6\u5167\u5bb9\u3001\u7248\u9762\u548c\u7d50\u69cb\u4e00\u4f75\u8b80\u51fa\u4f86\u3002\u5b83\u6700\u5be6\u969b\u7684\u50f9\u503c\uff0c\u5728\u65bc\u540c\u6642\u61c9\u4ed8 digital documents \u548c camera-captured documents\uff0c\u5c24\u5176\u9069\u5408\u6587\u4ef6\u62cd\u651d\u89d2\u5ea6\u4e0d\u6b63\u3001\u900f\u8996\u8b8a\u5f62\u6216\u7248\u9762\u8f03\u8907\u96dc\u7684\u60c5\u5883\u3002<\/p>\n\n\n\n<p class=\"wp-block-paragraph\">\u9019\u500b\u9805\u76ee\u5df2\u91cb\u51fa\u6a21\u578b\u6b0a\u91cd\u8207 technical report\uff0c\u4f7f\u7528\u65b9\u5f0f\u504f\u5411\u76f4\u63a5\u8f09\u5165 Hugging Face \u4e0a\u7684\u6a21\u578b\u505a\u63a8\u7406\u6216\u518d\u8a13\u7df4\uff0c\u800c\u4e0d\u662f\u4e00\u5957\u7368\u7acb\u61c9\u7528\u7a0b\u5f0f\u3002\u5b83\u7684\u8a13\u7df4\u6d41\u7a0b\u7d50\u5408 Multi-node Consensus Voting\uff08MCV\uff09\u3001Image-to-Image Self-Verification\uff0c\u4ee5\u53ca progressive four-stage training\uff0c\u986f\u793a\u4f5c\u8005\u628a\u5927\u91cf\u6a19\u8a3b\u6210\u672c\u8f49\u79fb\u5230\u81ea\u52d5\u5316\u8cc7\u6599\u6574\u7406\u8207\u9a57\u8b49\u3002<\/p>\n\n\n\n<ul class=\"wp-block-list\">\n<li>1.2B \u53c3\u6578\uff0c\u5b9a\u4f4d\u662f\u8f15\u91cf\u90e8\u7f72\u800c\u975e\u5806\u5927\u6a21\u578b\u3002<\/li>\n\n\n\n<li>\u540c\u6642\u9762\u5411\u6383\u63cf\u6587\u4ef6\u8207\u76f8\u6a5f\u62cd\u651d\u6587\u4ef6\uff0c\u8986\u84cb\u7bc4\u570d\u6bd4\u53ea\u505a\u55ae\u4e00\u8def\u5f91\u7684\u505a\u6cd5\u66f4\u5ee3\u3002<\/li>\n\n\n\n<li>Geometry-aware Document Modeling \u8207 Curvature-Guided Douglas-Peucker Sampling\uff08CGDP\uff09\u4e3b\u8981\u91dd\u5c0d\u5f4e\u66f2\u3001\u50be\u659c\u548c\u8b8a\u5f62\u9801\u9762\u3002<\/li>\n\n\n\n<li>Content-Structure Decoupled Learning \u4ee4\u8868\u683c\u8207\u516c\u5f0f\u7684\u5167\u5bb9\u548c\u7d50\u69cb\u5206\u958b\u5b78\uff0c\u8f03\u9069\u5408\u8907\u96dc\u6587\u4ef6\u3002<\/li>\n\n\n\n<li>\u4f5c\u8005\u8072\u7a31\u5728\u591a\u500b benchmark \u4e0a\u6709\u5f37\u8868\u73fe\uff0c\u4f46\u5177\u9ad4\u90e8\u7f72\u6210\u672c\u4ecd\u8981\u8996\u63a8\u7406\u6846\u67b6\u8207\u786c\u4ef6\u800c\u5b9a\u3002<\/li>\n<\/ul>\n\n\n\n<p class=\"wp-block-paragraph\">\u548c\u4e0d\u5c11\u53ea\u504f\u91cd OCR \u6587\u5b57\u62bd\u53d6\u7684\u505a\u6cd5\u76f8\u6bd4\uff0cNaviDC-OCR \u66f4\u50cf\u662f\u628a\u300c\u770b\u61c2\u6587\u4ef6\u300d\u653e\u5728\u7b2c\u4e00\u4f4d\uff1a\u5b83\u4e0d\u53ea\u8981\u8b80\u5b57\uff0c\u9084\u8981\u4fdd\u7559\u7248\u9762\u95dc\u4fc2\u8207\u5e7e\u4f55\u8cc7\u8a0a\u3002\u9019\u6703\u4ee4\u5b83\u5728\u4f01\u696d\u6587\u4ef6\u8655\u7406\u3001\u7968\u64da\u6574\u7406\u3001\u8868\u683c\u62bd\u53d6\u3001\u5b78\u8853\u6587\u4ef6\u89e3\u6790\uff0c\u6216\u8005\u624b\u6a5f\u62cd\u7167\u6383\u63cf\u7684\u5de5\u4f5c\u6d41\u88e1\u66f4\u6709\u7528\uff0c\u4f46\u5b83\u4ecd\u7136\u662f\u7814\u7a76\u578b\u6a21\u578b\uff0c\u843d\u5730\u6642\u8981\u7559\u610f\u901f\u5ea6\u3001\u8a18\u61b6\u9ad4\u8207\u5be6\u969b\u6587\u4ef6\u5206\u4f48\u662f\u5426\u63a5\u8fd1\u8a13\u7df4\u8cc7\u6599\u3002<\/p>\n\n\n\n<p class=\"wp-block-paragraph\"><a href=\"https:\/\/github.com\/caipeng328\/NaviDC-OCR\" rel=\"noopener noreferrer\"><strong>GitHub<\/strong><\/a> \u00b7 <a href=\"https:\/\/huggingface.co\/StarDoc-AI\/NaviDC-OCR\" rel=\"noopener noreferrer\"><strong>\u6a21\u578b<\/strong><\/a><\/p>\n","protected":false},"excerpt":{"rendered":"<p>NaviDC-OCR \u628a\u6587\u4ef6\u7406\u89e3\u6536\u6582\u5230\u4e00\u500b\u8f15\u91cf Vision-Language Model\uff08VLM\uff09\u6d41\u7a0b\uff0c\u5c0d\u6383\u63cf\u4ef6\u548c\u624b\u6a5f\u62cd\u651d\u6587\u4ef6\u90fd\u80fd\u4e00\u8d77\u8655\u7406\u3002\u5b83\u7684\u91cd\u9ede\u4e0d\u53ea\u5728\u8b58\u5225\u6587\u5b57\uff0c\u9084\u5728\u7248\u9762\u3001\u8868\u683c\u8207\u516c\u5f0f\u4e4b\u9593\u7684\u7d50\u69cb\u9084\u539f\u3002<\/p>\n","protected":false},"author":8,"featured_media":11149,"comment_status":"open","ping_status":"open","sticky":false,"template":"","format":"standard","meta":{"ai_generated_summary":"","footnotes":""},"categories":[133,185,30,119,127,149,199],"tags":[],"class_list":["post-11150","post","type-post","status-publish","format-standard","hentry","category-133","category-qwen","category-image","category-119","category-127","category-149","category-dataset-"],"_links":{"self":[{"href":"https:\/\/infernews.com\/blog\/wp-json\/wp\/v2\/posts\/11150","targetHints":{"allow":["GET"]}}],"collection":[{"href":"https:\/\/infernews.com\/blog\/wp-json\/wp\/v2\/posts"}],"about":[{"href":"https:\/\/infernews.com\/blog\/wp-json\/wp\/v2\/types\/post"}],"author":[{"embeddable":true,"href":"https:\/\/infernews.com\/blog\/wp-json\/wp\/v2\/users\/8"}],"replies":[{"embeddable":true,"href":"https:\/\/infernews.com\/blog\/wp-json\/wp\/v2\/comments?post=11150"}],"version-history":[{"count":1,"href":"https:\/\/infernews.com\/blog\/wp-json\/wp\/v2\/posts\/11150\/revisions"}],"predecessor-version":[{"id":11151,"href":"https:\/\/infernews.com\/blog\/wp-json\/wp\/v2\/posts\/11150\/revisions\/11151"}],"wp:featuredmedia":[{"embeddable":true,"href":"https:\/\/infernews.com\/blog\/wp-json\/wp\/v2\/media\/11149"}],"wp:attachment":[{"href":"https:\/\/infernews.com\/blog\/wp-json\/wp\/v2\/media?parent=11150"}],"wp:term":[{"taxonomy":"category","embeddable":true,"href":"https:\/\/infernews.com\/blog\/wp-json\/wp\/v2\/categories?post=11150"},{"taxonomy":"post_tag","embeddable":true,"href":"https:\/\/infernews.com\/blog\/wp-json\/wp\/v2\/tags?post=11150"}],"curies":[{"name":"wp","href":"https:\/\/api.w.org\/{rel}","templated":true}]}}