{"id":31662,"date":"2025-10-21T14:49:25","date_gmt":"2025-10-21T14:49:25","guid":{"rendered":"https:\/\/metaverseplanet.net\/blog\/?p=31662"},"modified":"2026-01-05T09:00:31","modified_gmt":"2026-01-05T09:00:31","slug":"deepseek-ocr-ai-doesnt-just-read-texts-it-sees-them","status":"publish","type":"post","link":"https:\/\/metaverseplanet.net\/blog\/deepseek-ocr-ai-doesnt-just-read-texts-it-sees-them\/","title":{"rendered":"Deepseek OCR: AI Doesn’t Just Read Texts, It “Sees” Them"},"content":{"rendered":"\n

Deepseek’s new OCR system processes texts as images and compresses them up to 10 times<\/strong>. This technology, capable of analyzing 33 million pages in a day, allows AI to read much longer documents.<\/p>\n\n\n\n

Deepseek, a Chinese artificial intelligence company, is attracting attention with its new OCR (Optical Character Recognition)<\/strong> system developed for more efficient processing of text-based documents. The system compresses image-based texts, enabling AI models to process much longer documents without hitting their memory limits.<\/p>\n\n\n\n

\n\n\n\n

Processing Text as Visual Data<\/strong><\/h2>\n\n\n\n
\"\"<\/figure>\n\n\n\n

According to Deepseek’s technical report, the system analyzes text data in image format instead of processing it directly. This approach significantly reduces the computational load. The new OCR system can compress texts by up to 10 times<\/strong> while retaining 97% of the information<\/strong>.<\/p>\n\n\n\n

As known, large language models represent text as tokens<\/strong>, with each token containing a few characters. Researchers are working to develop models that can process long documents and conversations exceeding millions of tokens, thereby expanding the context window<\/strong>. However, as the number of tokens that can be processed simultaneously increases, so do the computational costs. Thus, a large token capacity prevents the model’s memory from filling up even with long documents, but it increases the cost. Deepseek’s OCR solution, however, processes very long content as if it were an image<\/strong>, effectively viewing the content as pixels.<\/p>\n\n\n\n

\n\n\n\n

Seeing Long Texts as Pixels<\/strong><\/h2>\n\n\n\n
\"\"<\/figure>\n\n\n\n

The core of the system consists of two main components: DeepEncoder<\/strong> and Deepseek3B-MoE<\/strong>. DeepEncoder, which handles the image processing, operates with 380 million parameters. Deepseek3B-MoE, responsible for text generation, has 570 million active parameters. DeepEncoder combines Meta’s 80-million-parameter SAM (Segment Anything Model)<\/strong> and OpenAI’s 300-million-parameter CLIP model<\/strong>. An intermediary 16x compressor<\/strong> significantly reduces the image data, increasing processing speed. For example, 4,096 tokens of a $1,024 \\times 1,024$ pixel image are reduced to only 256 tokens<\/strong> after compression.<\/p>\n\n\n\n

Deepseek OCR can operate using between 64 and 400 “vision tokens<\/strong>,” depending on the resolution. This number significantly lightens operations that typically require thousands of tokens in classic OCR systems. In OmniDocBench tests, the system outperformed GOT-OCR 2.0 using only 100 vision tokens. It also surpassed the performance of MinerU 2.0, which required over 6,000 tokens, while operating under 800 tokens.<\/p>\n\n\n\n