🇨🇳 中文
围绕 PP-OCRv6 搭建的本地 OCR 工作台。PP-OCRv6 是飞桨最新的三档 OCR 模型家族（Tiny / Small / Medium），三档模型可以在本机一键切换，同时支持 OmniDocBench 标准评测集的本地跑分。
测试环境，macOS Apple Silicon（M 系列芯片）。ONNX Runtime 会自动启用 CoreML 加速，不需要额外配置。
拖拽上传区，支持批量处理和剪贴板粘贴（⌘V 截图直接粘）。
39 条历史记录，每张卡片显示缩略图、文件名、识别框数量和处理耗时。
报纸页，共识别出 193 个文本框。右侧面板展示完整文字转录结果。
模型档位切换（Tiny / Small / Medium）、CoreML 开关、检测参数滑块。
| 组件 | 说明 |
|---|---|
webapp/ |
FastAPI 后端 + 单页 Web UI。支持图片上传、模型切换、结果导出（CSV / Markdown / Excel）。 |
ppocrv6_browser.html |
零依赖浏览器 Demo，PP-OCRv6 Tiny 通过 ONNX Runtime Web 完全在浏览器内运行，无需服务器。 |
bench_local_v2.py |
通过本地 API 运行 OmniDocBench 评测。 |
run_apple_vision.py |
用 macOS 内置 Apple Vision 对同一批 18 张图跑分对照。 |
gen_result_vis.py |
为任意图片生成检测+识别结果可视化面板。 |
assets/realworld_ocr/ |
四张真实场景测试图 + PP-OCRv6 Medium 输出面板。 |
在 OmniDocBench 演示集（18 张文档页）上评测，指标，text_block 编辑距离，越低越好。
| 模型 | 文本块编辑距离 ↓ | 备注 |
|---|---|---|
| PP-OCRv6 Medium（34.5 MB） | 0.425 | 精度最高，Apple Silicon 本地运行 |
| PP-OCRv6 Small（7.7 MB） | 0.443 | 性能均衡，移动端体量 |
| PP-OCRv6 Tiny（1.5 MB） | 0.446 | 可在浏览器内跑，无需服务器 |
| Apple Vision（系统内置） | 0.448 | 零配置，0.16–0.54 秒/张 |
| PaddleOCR-VL（云端 API） | ~0.38* | 多模态，云端延迟 6–16 秒/张 |
*PaddleOCR-VL 单独评测，针对文档版面优化，非孤立文本块场景，数字仅供参考。
四张超出标准文档扫描范围的测试图，分别覆盖透视变形、点阵字体、浮雕低对比度文字、七段数码管四个场景。
斜拍加彩色底，透视角度让字体变形。Medium 完整读出，DESIGN SOUL / JONATHAN DOE / Creative Designer / 电话 / 网址。
两行文字均准确识别，DotMatrix（标题）+ RETRO PRINT CHARM（副标题）。字符集覆盖方面 Small 表现最稳。
曲面金属上的浮雕字，约 30° 斜角拍摄。Medium 读出，TREADWEAR 220 / PLACARD IN VEHICLE / TO SEAT BEADS / ME AXLE。这是最难的一个场景，Apple Vision 只读出了「220」，大多数多模态模型在低对比度下都吃力。
四块面板的产品编号（BVY413HSW、BVY411HSW）、品牌名（ORB ELEVATOR）、网址（orbelevator.en.alibaba.com）均正确识别。
| 最低 | 推荐 | |
|---|---|---|
| 操作系统 | macOS 13 Ventura | macOS 14+ Sonoma / Sequoia |
| 芯片 | Apple M1 | Apple M2 / M3 / M4 |
| 内存 | 8 GB | 16 GB |
| Python | 3.10 | 3.11 / 3.12 |
| 磁盘空间 | ~500 MB（模型 + 依赖） | — |
Intel Mac / Linux，使用标准
onnxruntime（仅 CPU）即可跑通。在webapp/server.py中去掉 CoreML Provider 相关代码，或在 UI 设置页把推理引擎改为 CPU。
git clone https://github.com/andyhuo520/ppocrv6-studio.git
cd ppocrv6-studiopython3 -m venv .venv
source .venv/bin/activatepip install -r requirements-webapp.txtApple Silicon 推荐额外安装 CoreML 加速版，
pip uninstall onnxruntime -y
pip install onnxruntime-siliconbash scripts/download_models.sh all脚本从 GitHub Releases 页面下载三档模型压缩包并解压，
ppocrv6_onnx/ ← Tiny（官方参数量 1.5 MB）
ppocrv6_small_onnx/ ← Small（7.7 MB）
ppocrv6_medium_onnx/ ← Medium（34.5 MB）
python webapp/server.py浏览器打开 **http://localhost:8765**。
PP-OCRv6 把 OCR 拆成两个阶段，
- 检测，用 DB（Differentiable Binarization）模型找出文字区域，骨干网络是 LCNetV4，感受野从 3×3 扩展到 7×7，对小字和密集排版的处理更稳。
- 识别，把检测到的区域逐个裁出来，用 CTC 模型加轻量注意力模块读字符。三档模型共用同一套骨干，只在宽度和深度上有差别。
ONNX Runtime 在 Apple Silicon 上把两个阶段都走 CoreML，实测单张耗时 3–52 秒，取决于档位和图片分辨率。
MIT。详见 LICENSE。
PP-OCRv6 模型权重由飞桨团队以 Apache 2.0 许可证 发布。
完整的评测思路、真实场景分析和横向对比，见 PP-OCRv6_Khazix.html（中文长文）。
🇺🇸 English
A local OCR workbench built around PP-OCRv6 — PaddlePaddle's latest three-tier OCR model family (Tiny / Small / Medium). Run all three tiers on your own machine, switch between them in one click, and benchmark against real-world edge cases and the OmniDocBench standard evaluation set.
Built and tested on macOS with Apple Silicon (M-series). CoreML acceleration is enabled automatically via ONNX Runtime.
Drag-and-drop upload zone. Supports batch processing and clipboard paste (⌘V).
39-item history grid. Each card shows the thumbnail, filename, box count, and processing time.
Newspaper page with 193 detection boxes. The right panel shows the full text transcript.
Model tier selector (Tiny / Small / Medium), CoreML toggle, and detection parameter sliders.
| Component | Description |
|---|---|
webapp/ |
FastAPI backend + single-page web UI. Upload images, switch models, export results as CSV / Markdown / Excel. |
ppocrv6_browser.html |
Zero-dependency browser demo — no server, runs PP-OCRv6 Tiny entirely in-browser via ONNX Runtime Web. |
bench_local_v2.py |
Run OmniDocBench evaluation against the local API server. |
run_apple_vision.py |
Benchmark Apple Vision Framework (macOS built-in) on the same 18-image set. |
gen_result_vis.py |
Generate side-by-side detection + recognition visualisation panels for arbitrary images. |
assets/realworld_ocr/ |
Four real-world test images + PP-OCRv6 Medium output panels. |
Evaluated on OmniDocBench demo set (18 document pages). Metric: text_block Edit Distance — lower is better.
| Model | Text Block ↓ | Notes |
|---|---|---|
| PP-OCRv6 Medium (34.5 MB) | 0.425 | Best accuracy; runs locally on Apple Silicon |
| PP-OCRv6 Small (7.7 MB) | 0.443 | Good balance; mobile-class size |
| PP-OCRv6 Tiny (1.5 MB) | 0.446 | Runs in browser via ONNX Runtime Web |
| Apple Vision (built-in) | 0.448 | Zero-setup; 0.16–0.54 s/image |
| PaddleOCR-VL (cloud API) | ~0.38* | Multimodal; cloud latency 6–16 s/image |
*PaddleOCR-VL evaluated separately; optimised for document layout, not isolated text blocks.
Four images that push OCR beyond clean document scans: perspective angles, dot-matrix fonts, embossed low-contrast text, and seven-segment LED displays.
Detection correctly locates all text regions despite the skewed angle and coloured background. Medium reads: DESIGN SOUL / JONATHAN DOE / Creative Designer / phone / url.
Both lines detected and recognised cleanly: DotMatrix (title) + RETRO PRINT CHARM (subtitle). Small model performs best here due to its character-set coverage.
Embossed text on curved metal at ~30° angle. Medium reads: TREADWEAR 220 / PLACARD IN VEHICLE / TO SEAT BEADS / ME AXLE. Hardest case — Apple Vision reads only "220", most multimodal models struggle with the low contrast.
All product codes (BVY413HSW, BVY411HSW), brand name (ORB ELEVATOR), and URL (orbelevator.en.alibaba.com) correctly detected across four panels.
| Minimum | Recommended | |
|---|---|---|
| OS | macOS 13 Ventura | macOS 14+ Sonoma / Sequoia |
| CPU | Apple M1 | Apple M2 / M3 / M4 |
| RAM | 8 GB | 16 GB |
| Python | 3.10 | 3.11 / 3.12 |
| Disk | ~500 MB (models + deps) | — |
Intel Mac / Linux: Works with
onnxruntime(CPU only). Remove the CoreML provider lines inwebapp/server.pyor setprovider: cpuin the UI.
git clone https://github.com/andyhuo520/ppocrv6-studio.git
cd ppocrv6-studiopython3 -m venv .venv
source .venv/bin/activatepip install -r requirements-webapp.txtFor Apple Silicon CoreML acceleration (recommended):
pip uninstall onnxruntime -y
pip install onnxruntime-siliconbash scripts/download_models.sh allThis downloads three tarballs from the GitHub Releases page and extracts them:
ppocrv6_onnx/ ← Tiny (1.5 MB official params)
ppocrv6_small_onnx/ ← Small (7.7 MB)
ppocrv6_medium_onnx/ ← Medium (34.5 MB)
python webapp/server.pyOpen http://localhost:8765 in your browser.
PP-OCRv6 splits OCR into two stages:
- Detection — finds text regions using a DB (Differentiable Binarization) model with LCNetV4 backbone. Receptive field expanded from 3×3 to 7×7 for better small-text and dense-text handling.
- Recognition — crops each detected region and reads characters using a CTC model with a lightweight attention module. One backbone serves all three tiers (Tiny / Small / Medium differ only in width and depth).
ONNX Runtime on Apple Silicon routes both stages through CoreML, giving 3–52 s/image depending on model tier and image resolution.
MIT. See LICENSE.
PP-OCRv6 model weights are released under the Apache 2.0 license by PaddlePaddle.
Built as part of a hands-on benchmark series — see the full write-up in PP-OCRv6_Khazix.html (Chinese).