Unverified90% confidenceFactTime unknown
Gemma 4支持完全离线状态下的图像文字提取（OCR），采用视觉-语言多模态架构
1
Sources
90%
Confidence
Long-term
Relevance
6/1/2026
First Seen
Sources
Related Entities
Related Claims
UnverifiedThe Gemma series of models uses the SigLIP vision encoder to process image inputs.75% similarUnverifiedGemma 4 E2B supports three major offline multimodal capabilities: AI Chat (offline conversation), Ask Image (photo recognition), and Ask Audio (local audio transcription).72% similarUnverifiedThe multimodal conversational system supports three categories of image understanding capabilities: image content recognition, text extraction (OCR), and image-text question answering.67% similarUnverified该工具集成了OCR光学字符识别功能，能够自动识别屏幕上的文字内容65% similarUnverifiedThe visual model receives each screenshot and outputs structured information, including OCR-recognized text, page layout, pixel coordinate bounding boxes of UI controls, and element states.64% similar
Cite This Claim
Stable URI
https://kongchang.com/claim/23873API
curl https://kongchang.com/api/v1/knowledge/claims/23873MCP
get_claim(id=23873)