[{"data":1,"prerenderedAt":-1},["ShallowReactive",2],{"similar-neuroneural--brainchop":3,"tool-neuroneural--brainchop":61},[4,18,26,36,44,53],{"id":5,"name":6,"github_repo":7,"description_zh":8,"stars":9,"difficulty_score":10,"last_commit_at":11,"category_tags":12,"status":17},4358,"openclaw","openclaw\u002Fopenclaw","OpenClaw 是一款专为个人打造的本地化 AI 助手，旨在让你在自己的设备上拥有完全可控的智能伙伴。它打破了传统 AI 助手局限于特定网页或应用的束缚，能够直接接入你日常使用的各类通讯渠道，包括微信、WhatsApp、Telegram、Discord、iMessage 等数十种平台。无论你在哪个聊天软件中发送消息，OpenClaw 都能即时响应，甚至支持在 macOS、iOS 和 Android 设备上进行语音交互，并提供实时的画布渲染功能供你操控。\n\n这款工具主要解决了用户对数据隐私、响应速度以及“始终在线”体验的需求。通过将 AI 部署在本地，用户无需依赖云端服务即可享受快速、私密的智能辅助，真正实现了“你的数据，你做主”。其独特的技术亮点在于强大的网关架构，将控制平面与核心助手分离，确保跨平台通信的流畅性与扩展性。\n\nOpenClaw 非常适合希望构建个性化工作流的技术爱好者、开发者，以及注重隐私保护且不愿被单一生态绑定的普通用户。只要具备基础的终端操作能力（支持 macOS、Linux 及 Windows WSL2），即可通过简单的命令行引导完成部署。如果你渴望拥有一个懂你",349277,3,"2026-04-06T06:32:30",[13,14,15,16],"Agent","开发框架","图像","数据工具","ready",{"id":19,"name":20,"github_repo":21,"description_zh":22,"stars":23,"difficulty_score":10,"last_commit_at":24,"category_tags":25,"status":17},3808,"stable-diffusion-webui","AUTOMATIC1111\u002Fstable-diffusion-webui","stable-diffusion-webui 是一个基于 Gradio 构建的网页版操作界面，旨在让用户能够轻松地在本地运行和使用强大的 Stable Diffusion 图像生成模型。它解决了原始模型依赖命令行、操作门槛高且功能分散的痛点，将复杂的 AI 绘图流程整合进一个直观易用的图形化平台。\n\n无论是希望快速上手的普通创作者、需要精细控制画面细节的设计师，还是想要深入探索模型潜力的开发者与研究人员，都能从中获益。其核心亮点在于极高的功能丰富度：不仅支持文生图、图生图、局部重绘（Inpainting）和外绘（Outpainting）等基础模式，还独创了注意力机制调整、提示词矩阵、负向提示词以及“高清修复”等高级功能。此外，它内置了 GFPGAN 和 CodeFormer 等人脸修复工具，支持多种神经网络放大算法，并允许用户通过插件系统无限扩展能力。即使是显存有限的设备，stable-diffusion-webui 也提供了相应的优化选项，让高质量的 AI 艺术创作变得触手可及。",162132,"2026-04-05T11:01:52",[14,15,13],{"id":27,"name":28,"github_repo":29,"description_zh":30,"stars":31,"difficulty_score":32,"last_commit_at":33,"category_tags":34,"status":17},1381,"everything-claude-code","affaan-m\u002Feverything-claude-code","everything-claude-code 是一套专为 AI 编程助手（如 Claude Code、Codex、Cursor 等）打造的高性能优化系统。它不仅仅是一组配置文件，而是一个经过长期实战打磨的完整框架，旨在解决 AI 代理在实际开发中面临的效率低下、记忆丢失、安全隐患及缺乏持续学习能力等核心痛点。\n\n通过引入技能模块化、直觉增强、记忆持久化机制以及内置的安全扫描功能，everything-claude-code 能显著提升 AI 在复杂任务中的表现，帮助开发者构建更稳定、更智能的生产级 AI 代理。其独特的“研究优先”开发理念和针对 Token 消耗的优化策略，使得模型响应更快、成本更低，同时有效防御潜在的攻击向量。\n\n这套工具特别适合软件开发者、AI 研究人员以及希望深度定制 AI 工作流的技术团队使用。无论您是在构建大型代码库，还是需要 AI 协助进行安全审计与自动化测试，everything-claude-code 都能提供强大的底层支持。作为一个曾荣获 Anthropic 黑客大奖的开源项目，它融合了多语言支持与丰富的实战钩子（hooks），让 AI 真正成长为懂上",143909,2,"2026-04-07T11:33:18",[14,13,35],"语言模型",{"id":37,"name":38,"github_repo":39,"description_zh":40,"stars":41,"difficulty_score":32,"last_commit_at":42,"category_tags":43,"status":17},2271,"ComfyUI","Comfy-Org\u002FComfyUI","ComfyUI 是一款功能强大且高度模块化的视觉 AI 引擎，专为设计和执行复杂的 Stable Diffusion 图像生成流程而打造。它摒弃了传统的代码编写模式，采用直观的节点式流程图界面，让用户通过连接不同的功能模块即可构建个性化的生成管线。\n\n这一设计巧妙解决了高级 AI 绘图工作流配置复杂、灵活性不足的痛点。用户无需具备编程背景，也能自由组合模型、调整参数并实时预览效果，轻松实现从基础文生图到多步骤高清修复等各类复杂任务。ComfyUI 拥有极佳的兼容性，不仅支持 Windows、macOS 和 Linux 全平台，还广泛适配 NVIDIA、AMD、Intel 及苹果 Silicon 等多种硬件架构，并率先支持 SDXL、Flux、SD3 等前沿模型。\n\n无论是希望深入探索算法潜力的研究人员和开发者，还是追求极致创作自由度的设计师与资深 AI 绘画爱好者，ComfyUI 都能提供强大的支持。其独特的模块化架构允许社区不断扩展新功能，使其成为当前最灵活、生态最丰富的开源扩散模型工具之一，帮助用户将创意高效转化为现实。",107888,"2026-04-06T11:32:50",[14,15,13],{"id":45,"name":46,"github_repo":47,"description_zh":48,"stars":49,"difficulty_score":32,"last_commit_at":50,"category_tags":51,"status":17},4721,"markitdown","microsoft\u002Fmarkitdown","MarkItDown 是一款由微软 AutoGen 团队打造的轻量级 Python 工具，专为将各类文件高效转换为 Markdown 格式而设计。它支持 PDF、Word、Excel、PPT、图片（含 OCR）、音频（含语音转录）、HTML 乃至 YouTube 链接等多种格式的解析，能够精准提取文档中的标题、列表、表格和链接等关键结构信息。\n\n在人工智能应用日益普及的今天，大语言模型（LLM）虽擅长处理文本，却难以直接读取复杂的二进制办公文档。MarkItDown 恰好解决了这一痛点，它将非结构化或半结构化的文件转化为模型“原生理解”且 Token 效率极高的 Markdown 格式，成为连接本地文件与 AI 分析 pipeline 的理想桥梁。此外，它还提供了 MCP（模型上下文协议）服务器，可无缝集成到 Claude Desktop 等 LLM 应用中。\n\n这款工具特别适合开发者、数据科学家及 AI 研究人员使用，尤其是那些需要构建文档检索增强生成（RAG）系统、进行批量文本分析或希望让 AI 助手直接“阅读”本地文件的用户。虽然生成的内容也具备一定可读性，但其核心优势在于为机器",93400,"2026-04-06T19:52:38",[52,14],"插件",{"id":54,"name":55,"github_repo":56,"description_zh":57,"stars":58,"difficulty_score":10,"last_commit_at":59,"category_tags":60,"status":17},4487,"LLMs-from-scratch","rasbt\u002FLLMs-from-scratch","LLMs-from-scratch 是一个基于 PyTorch 的开源教育项目，旨在引导用户从零开始一步步构建一个类似 ChatGPT 的大型语言模型（LLM）。它不仅是同名技术著作的官方代码库，更提供了一套完整的实践方案，涵盖模型开发、预训练及微调的全过程。\n\n该项目主要解决了大模型领域“黑盒化”的学习痛点。许多开发者虽能调用现成模型，却难以深入理解其内部架构与训练机制。通过亲手编写每一行核心代码，用户能够透彻掌握 Transformer 架构、注意力机制等关键原理，从而真正理解大模型是如何“思考”的。此外，项目还包含了加载大型预训练权重进行微调的代码，帮助用户将理论知识延伸至实际应用。\n\nLLMs-from-scratch 特别适合希望深入底层原理的 AI 开发者、研究人员以及计算机专业的学生。对于不满足于仅使用 API，而是渴望探究模型构建细节的技术人员而言，这是极佳的学习资源。其独特的技术亮点在于“循序渐进”的教学设计：将复杂的系统工程拆解为清晰的步骤，配合详细的图表与示例，让构建一个虽小但功能完备的大模型变得触手可及。无论你是想夯实理论基础，还是为未来研发更大规模的模型做准备",90106,"2026-04-06T11:19:32",[35,15,13,14],{"id":62,"github_repo":63,"name":64,"description_en":65,"description_zh":66,"ai_summary_zh":66,"readme_en":67,"readme_zh":68,"quickstart_zh":69,"use_case_zh":70,"hero_image_url":71,"owner_login":72,"owner_name":72,"owner_avatar_url":73,"owner_bio":74,"owner_company":75,"owner_location":75,"owner_email":75,"owner_twitter":75,"owner_website":75,"owner_url":76,"languages":77,"stars":105,"forks":106,"last_commit_at":107,"license":108,"difficulty_score":109,"env_os":110,"env_gpu":111,"env_ram":110,"env_deps":112,"category_tags":119,"github_topics":121,"view_count":32,"oss_zip_url":75,"oss_zip_packed_at":75,"status":17,"created_at":133,"updated_at":134,"faqs":135,"releases":170},5020,"neuroneural\u002Fbrainchop","brainchop","Brainchop: In-browser 3D MRI rendering and segmentation","Brainchop 是一款开源的神经影像前端工具，专为在网页浏览器中直接进行 3D MRI（磁共振成像）数据的可视化与自动分割而设计。它解决了传统医学影像处理依赖昂贵本地软件、复杂安装流程以及数据隐私担忧等痛点，让用户无需下载任何程序，只需打开网页即可上传并分析脑部扫描数据。\n\n该工具非常适合神经科学研究人员、放射科医生、医学生以及关注数据隐私的临床工作者使用。对于开发者而言，Brainchop 提供了纯 JavaScript 代码和预训练模型转换示例，便于二次开发或集成到现有系统中。\n\n其核心技术亮点在于利用轻量级深度学习模型（如 MeshNet），结合 TensorFlow.js 技术，将原本需要在服务器端运行的推理过程完全迁移至用户浏览器本地执行。这意味着所有敏感的医疗数据都在用户设备上处理，不会上传至云端，极大提升了安全性。同时，Brainchop 集成了 NiiVue 查看器，提供流畅的端到端操作体验，让复杂的 3D 脑部分割变得简单直观。无论是用于教学演示、快速原型验证，还是实际的科研分析，Brainchop 都为神经影像领域提供了一种便捷、安全且高效的解决方案。","\n# Brainchop  [![Version](https:\u002F\u002Fimg.shields.io\u002Fbadge\u002FVersion-4.0.0-brightgreen)]() [![JS ](https:\u002F\u002Fimg.shields.io\u002Fbadge\u002FTypes-JavaScript-blue)]() [![MIT-License ](https:\u002F\u002Fimg.shields.io\u002Fbadge\u002Flicense-MIT-green)](https:\u002F\u002Fgithub.com\u002Fneuroneural\u002Fbrainchop\u002Fblob\u002Fmaster\u002FLICENSE) [![tfjs](https:\u002F\u002Fimg.shields.io\u002Fbadge\u002Ftfjs-Pre--trained%20Model-blue)](https:\u002F\u002Fgithub.com\u002Fneuroneural\u002Fbrainchop\u002Ftree\u002Fmaster\u002Fmodels\u002Fmnm_tfjs_me_test) [![DOI](https:\u002F\u002Fjoss.theoj.org\u002Fpapers\u002F10.21105\u002Fjoss.05098\u002Fstatus.svg)](https:\u002F\u002Fdoi.org\u002F10.21105\u002Fjoss.05098)\n\n\n\u003Cdiv align=\"center\">\n  \u003Ca href=\"https:\u002F\u002Fneuroneural.github.io\u002Fbrainchop\">\n    \u003Cimg width=\"100%\" src=\"https:\u002F\u002Foss.gittoolsai.com\u002Fimages\u002Fneuroneural_brainchop_readme_1277bd1fb8bc.png\">\n  \u003C\u002Fa>\n\n\n**Frontend For Neuroimaging.  Open Source**\n\n**[brainchop.org](https:\u002F\u002Fneuroneural.github.io\u002Fbrainchop)    [Updates](#Updates)   [Doc](https:\u002F\u002Fgithub.com\u002Fneuroneural\u002Fbrainchop\u002Fwiki\u002F)   [News!](#News)   [Cite](#Citation)   [v3](https:\u002F\u002Fneuroneural.github.io\u002Fbrainchop\u002Fv3)**\n\n\u003C\u002Fdiv>\n\n\n\u003Cbr>\n \u003Cimg src=\"https:\u002F\u002Foss.gittoolsai.com\u002Fimages\u002Fneuroneural_brainchop_readme_921522e09a17.png\"  width=\"25%\" align=\"right\">\n\n \u003Cp align=\"justify\">\n \u003Cb>\u003Ca href=\"https:\u002F\u002Fneuroneural.github.io\u002Fbrainchop\u002F\"  style=\"text-decoration: none\"> Brainchop\u003C\u002Fa>\u003C\u002Fb> brings automatic 3D MRI  volumetric segmentation  capability to neuroimaging  by running a lightweight deep learning model (e.g., \u003Ca href=\"https:\u002F\u002Fmedium.com\u002Fpytorch\u002Fcatalyst-neuro-a-3d-brain-segmentation-pipeline-for-mri-b1bb1109276a\" target=\"_blank\"  style=\"text-decoration: none\"> MeshNet\u003C\u002Fa>) in the web-browser for inference on the user side. \n \u003C\u002Fp>\n\n \u003Cp align=\"justify\">\n We make the implementation of brainchop freely available, releasing its pure javascript code as open-source. The user interface (UI)  provides a web-based end-to-end solution for 3D MRI segmentation. \u003Cb>\u003Ca href=\"v\"  style=\"text-decoration: none\">NiiVue\u003C\u002Fa>\u003C\u002Fb> viewer is integrated with the tool for MRI visualization.  For more information about Brainchop, please refer to this detailed \u003Cb>\u003Ca href=\"https:\u002F\u002Fgithub.com\u002Fneuroneural\u002Fbrainchop\u002Fwiki\u002F\"  style=\"text-decoration: none\">Wiki\u003C\u002Fa>\u003C\u002Fb> and this \u003Cb>\u003Ca href=\"https:\u002F\u002Ftrendscenter.org\u002Fin-browser-3d-mri-segmentation-brainchop-org\u002F\"  style=\"text-decoration: none\"> Blog\u003C\u002Fa>\u003C\u002Fb>.\n\n  For questions or to share ideas, please refer to our  \u003Cb>\u003Ca href=\"https:\u002F\u002Fgithub.com\u002Fneuroneural\u002Fbrainchop\u002Fdiscussions\u002F\"  style=\"text-decoration: none\"> Discussions \u003C\u002Fa>\u003C\u002Fb> board.\n\n \u003C\u002Fp>\n\n\u003Cdiv align=\"center\">\n\n![Interface](https:\u002F\u002Foss.gittoolsai.com\u002Fimages\u002Fneuroneural_brainchop_readme_d8b34b8d7eeb.png)\n\n**Brainchop high-level architecture**\n\u003C\u002Fdiv>\n\n\n\u003Cdiv align=\"center\">\n\n![Interface](https:\u002F\u002Foss.gittoolsai.com\u002Fimages\u002Fneuroneural_brainchop_readme_7b8404861e88.png)\n\n**MeshNet deep learning architecture used for inference with Brainchop** (MeshNet  \u003Ca href=\"https:\u002F\u002Farxiv.org\u002Fpdf\u002F1612.00940.pdf\" target=\"_blank\"  style=\"text-decoration: none\"> paper\u003C\u002Fa>)\n\u003C\u002Fdiv>\n\n\n## MeshNet Example\nThis basic example provides an overview of the training pipeline for the MeshNet model. \n\n* [![Open In Colab](https:\u002F\u002Fcolab.research.google.com\u002Fassets\u002Fcolab-badge.svg)](https:\u002F\u002Fcolab.research.google.com\u002Fgithub\u002Fneuroneural\u002Fbrainchop\u002Fblob\u002Fmaster\u002Fpy2tfjs\u002FMeshNet_Training_Example.ipynb) [MeshNet basic training example](.\u002Fpy2tfjs\u002FMeshNet_Training_Example.ipynb)\n\n* [![Open In Colab](https:\u002F\u002Fcolab.research.google.com\u002Fassets\u002Fcolab-badge.svg)](https:\u002F\u002Fcolab.research.google.com\u002Fgithub\u002Fneuroneural\u002Fbrainchop\u002Fblob\u002Fmaster\u002Fpy2tfjs\u002FConvert_Trained_Model_To_TFJS.ipynb) [Convert the trained MeshNet model to tfjs model example ](.\u002Fpy2tfjs\u002FConvert_Trained_Model_To_TFJS.ipynb)\n\n\u003Cbr>\n\n## Live Demo\n\nTo see Brainchop **v4** in action please click  [here](https:\u002F\u002Fneuroneural.github.io\u002Fbrainchop). Or click on the gif below to see a video:\n\u003Cdiv align=\"center\">\n  \n[![Brainchop Overhaul](https:\u002F\u002Foss.gittoolsai.com\u002Fimages\u002Fneuroneural_brainchop_readme_1ed1cc4cd680.gif)](https:\u002F\u002Fgithub.com\u002Fneuroneural\u002Fbrainchop\u002Freleases\u002Fdownload\u002Fv4.1.0\u002FBrainchop_overhaul.mp4)\n\u003C\u002Fdiv>\n\nFor **v3** click [here](https:\u002F\u002Fneuroneural.github.io\u002Fbrainchop\u002Fv3).\n\n\u003Cbr>\n\n\n\n## Updates\n\n\u003Cdiv align=\"center\">\n\n \u003Cimg src=\"https:\u002F\u002Foss.gittoolsai.com\u002Fimages\u002Fneuroneural_brainchop_readme_77b9d6c6c560.png\" width=\"100%\">\n\n**Brainchop \u003Ca href= \"https:\u002F\u002Fneuroneural.github.io\u002Fbrainchop\u002F\" target=\"_blank\"  style=\"text-decoration: none\"> v4 \u003C\u002Fa> with \u003Ca href= \"https:\u002F\u002Fgithub.com\u002Fniivue\u002Fniivue\" target=\"_blank\"  style=\"text-decoration: none\"> NiiVue\u003C\u002Fa> viewer**\n\u003C\u002Fdiv>\n\n\u003Cbr>\n\n\u003Cdiv align=\"center\">\n\n  \u003Cimg src=\"https:\u002F\u002Foss.gittoolsai.com\u002Fimages\u002Fneuroneural_brainchop_readme_a0c66461191b.gif\"  width=\"60%\">\n\n**Brainchop \u003Ca href= \"https:\u002F\u002Fneuroneural.github.io\u002Fbrainchop\u002Fv3\" target=\"_blank\"  style=\"text-decoration: none\"> v3 \u003C\u002Fa> with more robust models**\n\u003C\u002Fdiv>\n\n\u003Cbr>\n\n\n\u003Cdiv align=\"center\">\n\n![Interface](https:\u002F\u002Foss.gittoolsai.com\u002Fimages\u002Fneuroneural_brainchop_readme_ce2b6ac6f810.gif)\n\n**Brainchop \u003Ca href= \"https:\u002F\u002Fneuroneural.github.io\u002Fbrainchop\u002Fv3\" target=\"_blank\"  style=\"text-decoration: none\"> v1.4.0 - v3.4.0 \u003C\u002Fa> rendering MRI Nifti file in 3D**\n\u003C\u002Fdiv>\n\n\u003Cbr>\n\n\u003Cdiv align=\"center\">\n\n![Interface](https:\u002F\u002Foss.gittoolsai.com\u002Fimages\u002Fneuroneural_brainchop_readme_8840a8c3335a.gif)\n\n\n**Brainchop \u003Ca href= \"https:\u002F\u002Fneuroneural.github.io\u002Fbrainchop\u002Fv3\" target=\"_blank\"  style=\"text-decoration: none\"> v1.3.0 - v3.4.0 \u003C\u002Fa>  rendering segmentation output in 3D**\n\u003C\u002Fdiv>\n\n\n\n\n\n## News!\n\n* Brainchop [v2.2.0](https:\u002F\u002Fgithub.com\u002Fneuroneural\u002Fbrainchop\u002Freleases\u002Ftag\u002Fv2.2.0) paper is accepted in the 21st IEEE International Symposium on Biomedical Imaging ([ISBI 2024](https:\u002F\u002Fbiomedicalimaging.org\u002F2024\u002F)). Lengthy arXiv version can be found [here](https:\u002F\u002Farxiv.org\u002Fabs\u002F2310.16162).\n\n\u003Cdiv align=\"center\">\n   \u003Cimg src=\"https:\u002F\u002Foss.gittoolsai.com\u002Fimages\u002Fneuroneural_brainchop_readme_b1189406b29a.jpeg\"  width=\"40%\">\n\u003C\u002Fdiv>\n\n\u003Cbr>\n\u003Cbr>\n\n* Brainchop [paper](https:\u002F\u002Fdoi.org\u002F10.21105\u002Fjoss.05098) is published in the Journal of Open Source Software (JOSS) on March 28, 2023.\n\n\u003Cdiv align=\"center\">\n   \u003Ca href=\"https:\u002F\u002Fdoi.org\u002F10.21105\u002Fjoss.05098\">\u003Cimg src=\"https:\u002F\u002Foss.gittoolsai.com\u002Fimages\u002Fneuroneural_brainchop_readme_cb0aca3c4a8f.png\">\u003C\u002Fa>\n\u003C\u002Fdiv>\n\n\u003Cbr>\n\u003Cbr>\n\n* Brainchop abstract is accepted for poster presentation during the 2023 [OHBM](https:\u002F\u002Fwww.humanbrainmapping.org\u002F) Annual Meeting.\n\n\u003Cdiv align=\"center\">\n   \u003Cimg src=\"https:\u002F\u002Foss.gittoolsai.com\u002Fimages\u002Fneuroneural_brainchop_readme_349cdd037b63.jpeg\"  width=\"40%\">\n\u003C\u002Fdiv>\n\n\u003Cbr>\n\u003Cbr>\n\n* Brainchop 1-page abstract and poster is accepted in 20th IEEE International Symposium on Biomedical Imaging ([ISBI 2023](https:\u002F\u002F2023.biomedicalimaging.org\u002Fen\u002F))\n\n\u003Cdiv align=\"center\">\n   \u003Cimg src=\"https:\u002F\u002Foss.gittoolsai.com\u002Fimages\u002Fneuroneural_brainchop_readme_c481f82f1165.png\"  width=\"40%\">\n\u003C\u002Fdiv>\n\n\u003Cbr>\n\u003Cbr>\n\n* Google, Tensorflow community spotlight award for brainchop (Sept 2022) on [Linkedin](https:\u002F\u002Fwww.linkedin.com\u002Fposts\u002Ftensorflow-community_github-neuroneuralbrainchop-brainchop-activity-6978796859532181504-cfCW?utm_source=share&utm_medium=member_desktop) and [Twitter](https:\u002F\u002Ftwitter.com\u002FTensorFlow\u002Fstatus\u002F1572980019999264774)\n\n\u003Cdiv align=\"center\">\n   \u003Cimg src=\"https:\u002F\u002Foss.gittoolsai.com\u002Fimages\u002Fneuroneural_brainchop_readme_de6656aa6522.png\"  width=\"60%\">\n\u003C\u002Fdiv>\n\n\u003Cbr>\n\u003Cbr>\n\n* Brainchop  invited to [Pytorch](https:\u002F\u002Fpytorch.org\u002Fecosystem\u002Fptc\u002F2022) flag conference, New Orleans, Louisiana (Dec 2022) \n\n\u003Cdiv align=\"center\">\n  \u003Cimg src=\"https:\u002F\u002Foss.gittoolsai.com\u002Fimages\u002Fneuroneural_brainchop_readme_bda1ff610da3.jpg\"  width=\"50%\">\n\u003C\u002Fdiv>\n\n\n\u003Cbr>\n\u003Cbr>\n\n* Brainchop  invited to TensorFlow.js Show & Tell episode #7 (Jul 2022). \n\n\u003Cdiv align=\"center\">\n  \u003Cimg src=\"https:\u002F\u002Foss.gittoolsai.com\u002Fimages\u002Fneuroneural_brainchop_readme_eb2ba974b1c5.png\"  width=\"50%\">\n\u003C\u002Fdiv>\n\n## Citation\n\nBrainchop [paper](https:\u002F\u002Fdoi.org\u002F10.21105\u002Fjoss.05098) for v2.1.0 is published on March 28, 2023, in the Journal of Open Source Software (JOSS) [![DOI](https:\u002F\u002Fjoss.theoj.org\u002Fpapers\u002F10.21105\u002Fjoss.05098\u002Fstatus.svg)](https:\u002F\u002Fdoi.org\u002F10.21105\u002Fjoss.05098) \n\n\n\u003Cbr>\n\nFor **APA** style, the paper can be **cited** as: \n\n> Masoud, M., Hu, F., & Plis, S. (2023). Brainchop: In-browser MRI volumetric segmentation and rendering. Journal of Open Source Software, 8(83), 5098. https:\u002F\u002Fdoi.org\u002F10.21105\u002Fjoss.05098\n\n\u003Cbr>\n\nFor **BibTeX** format that is used by some publishers,  please use: \n\n```BibTeX: \n@article{Masoud2023, \n  doi = {10.21105\u002Fjoss.05098}, \n  url = {https:\u002F\u002Fdoi.org\u002F10.21105\u002Fjoss.05098}, \n  year = {2023}, \n  publisher = {The Open Journal}, \n  volume = {8}, \n  number = {83}, \n  pages = {5098}, \n  author = {Mohamed Masoud and Farfalla Hu and Sergey Plis}, \n  title = {Brainchop: In-browser MRI volumetric segmentation and rendering}, \n  journal = {Journal of Open Source Software} \n} \n```\n\u003Cbr>\n\nFor **MLA** style: \n\n> Masoud, Mohamed, Farfalla Hu, and Sergey Plis. ‘Brainchop: In-Browser MRI Volumetric Segmentation and Rendering’. Journal of Open Source Software, vol. 8, no. 83, The Open Journal, 2023, p. 5098, https:\u002F\u002Fdoi.org10.21105\u002Fjoss.05098.\n\n\u003Cbr>\n\nFor **IEEE** style:\n\n> M. Masoud, F. Hu, and S. Plis, ‘Brainchop: In-browser MRI volumetric segmentation and rendering’, Journal of Open Source Software, vol. 8, no. 83, p. 5098, 2023. doi:10.21105\u002Fjoss.05098\n\n\n\u003Cbr>\n\n## Contribution and Authorship Guidelines\n\nIf you modify or extend Brainchop in a derivative work intended for publication (such as a research paper or software tool), please cite and acknowledge the original Brainchop project and the original authors. Proper acknowledge should include the following:\n\n> **\"Brainchop, originally developed by Mohamed Masoud and Sergey Plis (2023), was used in the development of this work.\"**\n\nWe also request that significant contributions to derivative works be recognized by including original authors as co-authors, where appropriate.\n\n\u003Cbr>\n\n## Funding\n\nThis work was funded by the NIH grant RF1MH121885. Additional support from NIH R01MH123610, R01EB006841 and NSF 2112455.\n\n\u003Cbr \u002F>\n\u003Cdiv align=\"center\">\n\n\u003Cimg src='https:\u002F\u002Foss.gittoolsai.com\u002Fimages\u002Fneuroneural_brainchop_readme_edb5023e2418.jpg' width='300' height='100'>\u003C\u002Fimg>\n\n**Mohamed Masoud - Sergey Plis - 2024**\n\u003C\u002Fdiv>\n","# Brainchop  [![版本](https:\u002F\u002Fimg.shields.io\u002Fbadge\u002FVersion-4.0.0-brightgreen)]() [![JS ](https:\u002F\u002Fimg.shields.io\u002Fbadge\u002FTypes-JavaScript-blue)]() [![MIT许可证](https:\u002F\u002Fimg.shields.io\u002Fbadge\u002Flicense-MIT-green)](https:\u002F\u002Fgithub.com\u002Fneuroneural\u002Fbrainchop\u002Fblob\u002Fmaster\u002FLICENSE) [![tfjs](https:\u002F\u002Fimg.shields.io\u002Fbadge\u002Ftfjs-预训练模型-blue)](https:\u002F\u002Fgithub.com\u002Fneuroneural\u002Fbrainchop\u002Ftree\u002Fmaster\u002Fmodels\u002Fmnm_tfjs_me_test) [![DOI](https:\u002F\u002Fjoss.theoj.org\u002Fpapers\u002F10.21105\u002Fjoss.05098\u002Fstatus.svg)](https:\u002F\u002Fdoi.org\u002F10.21105\u002Fjoss.05098)\n\n\n\u003Cdiv align=\"center\">\n  \u003Ca href=\"https:\u002F\u002Fneuroneural.github.io\u002Fbrainchop\">\n    \u003Cimg width=\"100%\" src=\"https:\u002F\u002Foss.gittoolsai.com\u002Fimages\u002Fneuroneural_brainchop_readme_1277bd1fb8bc.png\">\n  \u003C\u002Fa>\n\n\n**面向神经影像的前端工具。开源**\n\n**[brainchop.org](https:\u002F\u002Fneuroneural.github.io\u002Fbrainchop)    [更新](#Updates)   [文档](https:\u002F\u002Fgithub.com\u002Fneuroneural\u002Fbrainchop\u002Fwiki\u002F)   [新闻!](#News)   [引用](#Citation)   [v3](https:\u002F\u002Fneuroneural.github.io\u002Fbrainchop\u002Fv3)**\n\n\u003C\u002Fdiv>\n\n\n\u003Cbr>\n \u003Cimg src=\"https:\u002F\u002Foss.gittoolsai.com\u002Fimages\u002Fneuroneural_brainchop_readme_921522e09a17.png\"  width=\"25%\" align=\"right\">\n\n \u003Cp align=\"justify\">\n \u003Cb>\u003Ca href=\"https:\u002F\u002Fneuroneural.github.io\u002Fbrainchop\u002F\"  style=\"text-decoration: none\"> Brainchop\u003C\u002Fa>\u003C\u002Fb>通过在用户端的Web浏览器中运行轻量级深度学习模型（例如，\u003Ca href=\"https:\u002F\u002Fmedium.com\u002Fpytorch\u002Fcatalyst-neuro-a-3d-brain-segmentation-pipeline-for-mri-b1bb1109276a\" target=\"_blank\"  style=\"text-decoration: none\"> MeshNet\u003C\u002Fa>)进行推理，为神经影像学带来了自动化的3D MRI体积分割能力。 \n \u003C\u002Fp>\n\n \u003Cp align=\"justify\">\n 我们免费提供了Brainchop的实现，并将其纯JavaScript代码以开源形式发布。该用户界面（UI）提供了一个基于Web的端到端3D MRI分割解决方案。\u003Cb>\u003Ca href=\"v\"  style=\"text-decoration: none\">NiiVue\u003C\u002Fa>\u003C\u002Fb>查看器已集成到该工具中，用于MRI可视化。有关Brainchop的更多信息，请参阅这份详细的\u003Cb>\u003Ca href=\"https:\u002F\u002Fgithub.com\u002Fneuroneural\u002Fbrainchop\u002Fwiki\u002F\"  style=\"text-decoration: none\">维基\u003C\u002Fa>\u003C\u002Fb>以及这篇\u003Cb>\u003Ca href=\"https:\u002F\u002Ftrendscenter.org\u002Fin-browser-3d-mri-segmentation-brainchop-org\u002F\"  style=\"text-decoration: none\">博客文章\u003C\u002Fa>\u003C\u002Fb>。\n\n  如有任何问题或想法想要分享，请访问我们的\u003Cb>\u003Ca href=\"https:\u002F\u002Fgithub.com\u002Fneuroneural\u002Fbrainchop\u002Fdiscussions\u002F\"  style=\"text-decoration: none\">讨论区\u003C\u002Fa>\u003C\u002Fb>。\n\n \u003C\u002Fp>\n\n\u003Cdiv align=\"center\">\n\n![界面](https:\u002F\u002Foss.gittoolsai.com\u002Fimages\u002Fneuroneural_brainchop_readme_d8b34b8d7eeb.png)\n\n**Brainchop 高层次架构**\n\u003C\u002Fdiv>\n\n\n\u003Cdiv align=\"center\">\n\n![界面](https:\u002F\u002Foss.gittoolsai.com\u002Fimages\u002Fneuroneural_brainchop_readme_7b8404861e88.png)\n\n**Brainchop 中使用的 MeshNet 深度学习架构**（MeshNet  \u003Ca href=\"https:\u002F\u002Farxiv.org\u002Fpdf\u002F1612.00940.pdf\" target=\"_blank\"  style=\"text-decoration: none\">论文\u003C\u002Fa>)\n\u003C\u002Fdiv>\n\n\n## MeshNet 示例\n这个基础示例概述了MeshNet模型的训练流程。\n\n* [![在Colab中打开](https:\u002F\u002Fcolab.research.google.com\u002Fassets\u002Fcolab-badge.svg)](https:\u002F\u002Fcolab.research.google.com\u002Fgithub\u002Fneuroneural\u002Fbrainchop\u002Fblob\u002Fmaster\u002Fpy2tfjs\u002FMeshNet_Training_Example.ipynb) [MeshNet 基础训练示例](.\u002Fpy2tfjs\u002FMeshNet_Training_Example.ipynb)\n\n* [![在Colab中打开](https:\u002F\u002Fcolab.research.google.com\u002Fassets\u002Fcolab-badge.svg)](https:\u002F\u002Fcolab.research.google.com\u002Fgithub\u002Fneuroneural\u002Fbrainchop\u002Fblob\u002Fmaster\u002Fpy2tfjs\u002FConvert_Trained_Model_To_TFJS.ipynb) [将训练好的MeshNet模型转换为tfjs模型示例 ](.\u002Fpy2tfjs\u002FConvert_Trained_Model_To_TFJS.ipynb)\n\n\u003Cbr>\n\n## 实时演示\n\n要查看Brainchop **v4** 的实际效果，请点击  [这里](https:\u002F\u002Fneuroneural.github.io\u002Fbrainchop)。或者点击下方的动图观看视频：\n\u003Cdiv align=\"center\">\n  \n[![Brainchop 全新升级](https:\u002F\u002Foss.gittoolsai.com\u002Fimages\u002Fneuroneural_brainchop_readme_1ed1cc4cd680.gif)](https:\u002F\u002Fgithub.com\u002Fneuroneural\u002Fbrainchop\u002Freleases\u002Fdownload\u002Fv4.1.0\u002FBrainchop_overhaul.mp4)\n\u003C\u002Fdiv>\n\n如需查看 **v3**，请点击 [这里](https:\u002F\u002Fneuroneural.github.io\u002Fbrainchop\u002Fv3)。\n\n\u003Cbr>\n\n\n\n## 更新\n\n\u003Cdiv align=\"center\">\n\n \u003Cimg src=\"https:\u002F\u002Foss.gittoolsai.com\u002Fimages\u002Fneuroneural_brainchop_readme_77b9d6c6c560.png\" width=\"100%\">\n\n**Brainchop \u003Ca href= \"https:\u002F\u002Fneuroneural.github.io\u002Fbrainchop\u002F\" target=\"_blank\"  style=\"text-decoration: none\"> v4 \u003C\u002Fa> 搭配 \u003Ca href= \"https:\u002F\u002Fgithub.com\u002Fniivue\u002Fniivue\" target=\"_blank\"  style=\"text-decoration: none\"> NiiVue\u003C\u002Fa> 查看器**\n\u003C\u002Fdiv>\n\n\u003Cbr>\n\n\u003Cdiv align=\"center\">\n\n  \u003Cimg src=\"https:\u002F\u002Foss.gittoolsai.com\u002Fimages\u002Fneuroneural_brainchop_readme_a0c66461191b.gif\"  width=\"60%\">\n\n**Brainchop \u003Ca href= \"https:\u002F\u002Fneuroneural.github.io\u002Fbrainchop\u002Fv3\" target=\"_blank\"  style=\"text-decoration: none\"> v3 \u003C\u002Fa> 拥有更稳健的模型**\n\u003C\u002Fdiv>\n\n\u003Cbr>\n\n\n\u003Cdiv align=\"center\">\n\n![界面](https:\u002F\u002Foss.gittoolsai.com\u002Fimages\u002Fneuroneural_brainchop_readme_ce2b6ac6f810.gif)\n\n**Brainchop \u003Ca href= \"https:\u002F\u002Fneuroneural.github.io\u002Fbrainchop\u002Fv3\" target=\"_blank\"  style=\"text-decoration: none\"> v1.4.0 - v3.4.0 \u003C\u002Fa> 在3D中渲染MRI Nifti文件**\n\u003C\u002Fdiv>\n\n\u003Cbr>\n\n\u003Cdiv align=\"center\">\n\n![界面](https:\u002F\u002Foss.gittoolsai.com\u002Fimages\u002Fneuroneural_brainchop_readme_8840a8c3335a.gif)\n\n\n**Brainchop \u003Ca href= \"https:\u002F\u002Fneuroneural.github.io\u002Fbrainchop\u002Fv3\" target=\"_blank\"  style=\"text-decoration: none\"> v1.3.0 - v3.4.0 \u003C\u002Fa> 在3D中渲染分割结果**\n\u003C\u002Fdiv>\n\n## 新闻！\n\n* Brainchop [v2.2.0](https:\u002F\u002Fgithub.com\u002Fneuroneural\u002Fbrainchop\u002Freleases\u002Ftag\u002Fv2.2.0) 的论文已被第21届IEEE国际生物医学成像研讨会（[ISBI 2024](https:\u002F\u002Fbiomedicalimaging.org\u002F2024\u002F)）接收。更长的arXiv版本可在此处找到：[arXiv链接](https:\u002F\u002Farxiv.org\u002Fabs\u002F2310.16162)。\n\n\u003Cdiv align=\"center\">\n   \u003Cimg src=\"https:\u002F\u002Foss.gittoolsai.com\u002Fimages\u002Fneuroneural_brainchop_readme_b1189406b29a.jpeg\"  width=\"40%\">\n\u003C\u002Fdiv>\n\n\u003Cbr>\n\u003Cbr>\n\n* Brainchop [论文](https:\u002F\u002Fdoi.org\u002F10.21105\u002Fjoss.05098)于2023年3月28日发表在开源软件期刊（JOSS）上。\n\n\u003Cdiv align=\"center\">\n   \u003Ca href=\"https:\u002F\u002Fdoi.org\u002F10.21105\u002Fjoss.05098\">\u003Cimg src=\"https:\u002F\u002Foss.gittoolsai.com\u002Fimages\u002Fneuroneural_brainchop_readme_cb0aca3c4a8f.png\">\u003C\u002Fa>\n\u003C\u002Fdiv>\n\n\u003Cbr>\n\u003Cbr>\n\n* Brainchop的摘要被接受在2023年[OHBM](https:\u002F\u002Fwww.humanbrainmapping.org\u002F)年会期间以海报形式展示。\n\n\u003Cdiv align=\"center\">\n   \u003Cimg src=\"https:\u002F\u002Foss.gittoolsai.com\u002Fimages\u002Fneuroneural_brainchop_readme_349cdd037b63.jpeg\"  width=\"40%\">\n\u003C\u002Fdiv>\n\n\u003Cbr>\n\u003Cbr>\n\n* Brainchop的1页摘要和海报被第20届IEEE国际生物医学成像研讨会（[ISBI 2023](https:\u002F\u002F2023.biomedicalimaging.org\u002Fen\u002F)）接受。\n\n\u003Cdiv align=\"center\">\n   \u003Cimg src=\"https:\u002F\u002Foss.gittoolsai.com\u002Fimages\u002Fneuroneural_brainchop_readme_c481f82f1165.png\"  width=\"40%\">\n\u003C\u002Fdiv>\n\n\u003Cbr>\n\u003Cbr>\n\n* Google、TensorFlow社区为Brainchop颁发了聚光灯奖（2022年9月），相关信息分别发布在[LinkedIn](https:\u002F\u002Fwww.linkedin.com\u002Fposts\u002Ftensorflow-community_github-neuroneuralbrainchop-brainchop-activity-6978796859532181504-cfCW?utm_source=share&utm_medium=member_desktop)和[Twitter](https:\u002F\u002Ftwitter.com\u002FTensorFlow\u002Fstatus\u002F1572980019999264774)上。\n\n\u003Cdiv align=\"center\">\n   \u003Cimg src=\"https:\u002F\u002Foss.gittoolsai.com\u002Fimages\u002Fneuroneural_brainchop_readme_de6656aa6522.png\"  width=\"60%\">\n\u003C\u002Fdiv>\n\n\u003Cbr>\n\u003Cbr>\n\n* Brainchop受邀参加[PyTorch](https:\u002F\u002Fpytorch.org\u002Fecosystem\u002Fptc\u002F2022)旗帜会议，地点：路易斯安那州新奥尔良市（2022年12月）。\n\n\u003Cdiv align=\"center\">\n  \u003Cimg src=\"https:\u002F\u002Foss.gittoolsai.com\u002Fimages\u002Fneuroneural_brainchop_readme_bda1ff610da3.jpg\"  width=\"50%\">\n\u003C\u002Fdiv>\n\n\n\u003Cbr>\n\u003Cbr>\n\n* Brainchop受邀参加TensorFlow.js Show & Tell第七期节目（2022年7月）。\n\n\u003Cdiv align=\"center\">\n  \u003Cimg src=\"https:\u002F\u002Foss.gittoolsai.com\u002Fimages\u002Fneuroneural_brainchop_readme_eb2ba974b1c5.png\"  width=\"50%\">\n\u003C\u002Fdiv>\n\n## 引用\n\nBrainchop [论文](https:\u002F\u002Fdoi.org\u002F10.21105\u002Fjoss.05098)针对v2.1.0版本于2023年3月28日发表在开源软件期刊（JOSS）上 [![DOI](https:\u002F\u002Fjoss.theoj.org\u002Fpapers\u002F10.21105\u002Fjoss.05098\u002Fstatus.svg)](https:\u002F\u002Fdoi.org\u002F10.21105\u002Fjoss.05098) \n\n\n\u003Cbr>\n\n对于**APA**格式，该论文可被**引用**如下：\n\n> Masoud, M., Hu, F., & Plis, S. (2023). Brainchop: 在浏览器中进行MRI体积分割与渲染。开源软件期刊，8(83), 5098. https:\u002F\u002Fdoi.org\u002F10.21105\u002Fjoss.05098\n\n\u003Cbr>\n\n对于一些出版商使用的**BibTeX**格式，请使用以下内容：\n\n```BibTeX: \n@article{Masoud2023, \n  doi = {10.21105\u002Fjoss.05098}, \n  url = {https:\u002F\u002Fdoi.org\u002F10.21105\u002Fjoss.05098}, \n  year = {2023}, \n  publisher = {The Open Journal}, \n  volume = {8}, \n  number = {83}, \n  pages = {5098}, \n  author = {Mohamed Masoud and Farfalla Hu and Sergey Plis}, \n  title = {Brainchop: 在浏览器中进行MRI体积分割与渲染}, \n  journal = {开源软件期刊} \n} \n```\n\u003Cbr>\n\n对于**MLA**格式：\n\n> Masoud, Mohamed, Farfalla Hu, and Sergey Plis. ‘Brainchop: 在浏览器中进行MRI体积分割与渲染’. 开源软件期刊，第8卷，第83期，The Open Journal，2023年，第5098页，https:\u002F\u002Fdoi.org10.21105\u002Fjoss.05098.\n\n\u003Cbr>\n\n对于**IEEE**格式：\n\n> M. Masoud, F. Hu, 和 S. Plis，“Brainchop：在浏览器中进行MRI体积分割与渲染”，开源软件期刊，第8卷，第83期，第5098页，2023年。doi:10.21105\u002Fjoss.05098\n\n\n\u003Cbr>\n\n## 贡献与作者署名指南\n\n如果您对Brainchop进行修改或扩展，并将其用于拟发表的衍生作品（如研究论文或软件工具），请务必引用并致谢原始的Brainchop项目及其原作者。适当的致谢应包括以下内容：\n\n> **“Brainchop由Mohamed Masoud和Sergey Plis于2023年最初开发，在本研究工作中得到了应用。”**\n\n我们还建议，如果您的衍生作品中有重大贡献，应在适当的情况下将原作者列为共同作者。\n\n\u003Cbr>\n\n## 资助信息\n\n本研究工作得到了美国国立卫生研究院RF1MH121885项目的资助。此外，还获得了NIH R01MH123610、R01EB006841以及NSF 2112455的支持。\n\n\u003Cbr \u002F>\n\u003Cdiv align=\"center\">\n\n\u003Cimg src='https:\u002F\u002Foss.gittoolsai.com\u002Fimages\u002Fneuroneural_brainchop_readme_edb5023e2418.jpg' width='300' height='100'>\u003C\u002Fimg>\n\n**Mohamed Masoud - Sergey Plis - 2024**\n\u003C\u002Fdiv>","# Brainchop 快速上手指南\n\nBrainchop 是一个开源的前端神经影像工具，能够在浏览器中直接运行轻量级深度学习模型（如 MeshNet），实现 3D MRI 图像的自动体积分割。无需后端服务器，所有推理均在用户本地完成。\n\n## 环境准备\n\n由于 Brainchop 主要作为纯 JavaScript 应用在浏览器中运行，**普通用户无需配置复杂的本地开发环境**即可使用核心功能。\n\n*   **操作系统**：Windows、macOS、Linux 均可（仅需现代浏览器）。\n*   **浏览器要求**：推荐使用最新版本的 **Google Chrome**、**Microsoft Edge** 或 **Firefox**（需支持 WebGL 和 TensorFlow.js）。\n*   **输入数据格式**：支持标准的 NIfTI (`.nii`, `.nii.gz`) 格式的 3D MRI 图像。\n\n> **开发者注意**：如果您希望训练自定义模型或修改源码，则需要安装 Python 环境、PyTorch\u002FTensorFlow 以及 Node.js，并参考项目中的 `py2tfjs` 目录进行模型转换。\n\n## 安装步骤\n\nBrainchop 采用 **SaaS (Software as a Service)** 模式，**无需下载或安装任何软件**。\n\n1.  **访问官方演示站**：\n    直接在浏览器中打开以下地址即可启动最新版本 (v4)：\n    \n    ```text\n    https:\u002F\u002Fneuroneural.github.io\u002Fbrainchop\n    ```\n\n2.  **（可选）本地部署**：\n    如果您需要在内网环境使用或进行二次开发，可以通过克隆仓库并在本地启动静态服务器：\n\n    ```bash\n    # 克隆仓库\n    git clone https:\u002F\u002Fgithub.com\u002Fneuroneural\u002Fbrainchop.git\n    \n    # 进入目录\n    cd brainchop\n    \n    # 使用 Python 快速启动本地服务 (需安装 Python)\n    python -m http.server 8000\n    \n    # 或者使用 Node.js (需安装 http-server)\n    npx http-server -p 8000\n    ```\n    启动后，在浏览器访问 `http:\u002F\u002Flocalhost:8000`。\n\n## 基本使用\n\n以下是使用 Brainchop 进行 MRI 分割的最简流程：\n\n1.  **加载图像**：\n    *   点击界面上的 **\"Load Image\"** 按钮。\n    *   选择本地的 3D MRI NIfTI 文件（`.nii` 或 `.nii.gz`）。\n    *   系统将利用集成的 **NiiVue** 查看器在浏览器中渲染图像。\n\n2.  **执行分割**：\n    *   点击 **\"Run Segmentation\"**（或类似功能的开始按钮）。\n    *   浏览器将自动加载预训练的 **MeshNet** 模型并在本地进行推理。\n    *   *注意：首次运行可能需要几秒钟加载模型权重，后续处理速度取决于本地显卡性能。*\n\n3.  **查看与导出结果**：\n    *   分割完成后，界面将叠加显示分割掩膜（Mask），支持 3D 旋转查看。\n    *   点击 **\"Download\"** 或 **\"Save Mask\"** 按钮，将分割结果保存为新的 NIfTI 文件到本地。\n\n---\n*提示：本工具完全在客户端运行，您的医疗影像数据不会上传至任何服务器，确保了数据隐私安全。*","某神经科学实验室的研究员急需对一批新采集的脑部 MRI 数据进行初步分割，以评估患者海马体体积，但团队缺乏高性能 GPU 服务器且成员编程背景各异。\n\n### 没有 brainchop 时\n- **环境配置繁琐**：研究员需花费数小时在本地安装 Python、PyTorch 及各类依赖库，常因版本冲突导致运行失败。\n- **硬件门槛高**：3D 深度学习推理严重依赖昂贵的高端 GPU，普通办公笔记本无法承载，必须排队等待共享服务器资源。\n- **数据隐私风险**：若使用云端 API 处理敏感的患者医疗影像，需经过复杂的脱敏流程并面临数据外传合规性担忧。\n- **协作效率低下**：非技术背景的医生无法独立操作命令行工具，必须依赖工程师协助才能查看分割结果，沟通成本极高。\n\n### 使用 brainchop 后\n- **开箱即用**：研究员只需打开浏览器上传 NIfTI 格式文件，brainchop 即可自动加载预训练 MeshNet 模型开始推理，零配置启动。\n- **终端算力解放**：利用 TensorFlow.js 技术在用户端浏览器内直接运行轻量级模型，无需任何专用显卡，普通笔记本也能流畅处理。\n- **数据本地闭环**：所有计算均在浏览器沙箱中完成，原始医疗数据无需离开本地设备，彻底消除隐私泄露隐患。\n- **可视化即时反馈**：内置 NiiVue 查看器让医生能实时交互旋转 3D 脑图并调整分割阈值，实现了从上传到分析的一站式自助服务。\n\nbrainchop 通过将复杂的神经影像分割流程轻量化并迁移至浏览器，打破了硬件与技能壁垒，让高质量的 3D MRI 分析变得触手可及且安全合规。","https:\u002F\u002Foss.gittoolsai.com\u002Fimages\u002Fneuroneural_brainchop_ce2b6ac6.png","neuroneural","https:\u002F\u002Foss.gittoolsai.com\u002Favatars\u002Fneuroneural_45279daa.png","",null,"https:\u002F\u002Fgithub.com\u002Fneuroneural",[78,82,86,90,94,97,101],{"name":79,"color":80,"percentage":81},"JavaScript","#f1e05a",73.5,{"name":83,"color":84,"percentage":85},"CSS","#663399",14.9,{"name":87,"color":88,"percentage":89},"Jupyter Notebook","#DA5B0B",6.4,{"name":91,"color":92,"percentage":93},"SCSS","#c6538c",1.7,{"name":95,"color":96,"percentage":93},"Less","#1d365d",{"name":98,"color":99,"percentage":100},"HTML","#e34c26",1.3,{"name":102,"color":103,"percentage":104},"Python","#3572A5",0.6,524,62,"2026-04-07T06:11:50","MIT",1,"未说明","不需要专用 GPU（基于浏览器的推理，使用 TensorFlow.js）",{"notes":113,"python":114,"dependencies":115},"Brainchop 是一个纯前端工具，核心功能在 Web 浏览器中运行，无需安装本地深度学习环境。用户只需通过浏览器访问即可进行 3D MRI 分割。README 中提供的 Python 代码和 Colab 链接仅用于演示如何训练 MeshNet 模型并将其转换为 TensorFlow.js 格式，供开发者参考，普通用户运行工具时无需 Python 环境。","仅用于模型训练和转换示例（非运行时必需），版本未明确指定",[116,117,118],"TensorFlow.js (tfjs)","NiiVue","MeshNet (预训练模型)",[15,14,120],"其他",[122,123,124,125,126,127,128,129,130,131,132],"deep-learning","3d-segmentation","frontend-app","javascript","neuroimaging","pyodide","tensorflowjs","three-js","medical-imaging","mri","mri-segmentation","2026-03-27T02:49:30.150509","2026-04-07T22:51:04.212568",[136,141,146,150,155,160,165],{"id":137,"question_zh":138,"answer_zh":139,"source_url":140},22819,"Brainchop v3 是否支持输入 3D 图像并输出体积数据（如 Nifti 文件）？","是的，Brainchop v3 支持该功能。您可以像以前一样将分割结果保存为 Nifti 文件。操作方法是点击界面上的\"Save Overlay\"（保存覆盖层）按钮，然后在弹出的保存文件对话框中任意命名输出文件即可。","https:\u002F\u002Fgithub.com\u002Fneuroneural\u002Fbrainchop\u002Fissues\u002F40",{"id":142,"question_zh":143,"answer_zh":144,"source_url":145},22820,"如何在本地运行 Brainchop v3 版本？","要在本地运行 v3 版本，请进入项目目录下的 v3 文件夹，并通过命令行启动本地服务器。例如，若选择 8015 作为端口号，可执行以下命令：\n1. 进入目录：cd {您的本地路径}\u002Fbrainchop\u002Fv3\n2. 启动服务器：python -m http.server 8015\n3. 在浏览器中访问：http:\u002F\u002Flocalhost:8015\u002F","https:\u002F\u002Fgithub.com\u002Fneuroneural\u002Fbrainchop\u002Fissues\u002F55",{"id":147,"question_zh":148,"answer_zh":149,"source_url":145},22821,"Brainchop v3 是否支持直接处理 DICOM 格式文件？","对于 v3 版本，建议先将 DICOM 数据转换为 Nifti 格式后再使用。虽然在线版本的某些界面可能允许拖入 DICOM 文件查看，但为了获得最佳兼容性和处理流程，官方推荐进行格式转换。",{"id":151,"question_zh":152,"answer_zh":153,"source_url":154},22822,"Brainchop 模型是使用什么数据集训练的？","Brainchop 使用的是人类连接组计划（Human Connectome Project, HCP）数据集进行训练。这是一个公开访问的数据集，常用于脑部 MRI 分割研究。","https:\u002F\u002Fgithub.com\u002Fneuroneural\u002Fbrainchop\u002Fissues\u002F15",{"id":156,"question_zh":157,"answer_zh":158,"source_url":159},22823,"MeshNet 训练示例笔记本（Notebook）运行失败怎么办？","如果本地运行示例失败，建议使用官方提供的 Google Colab 链接，该环境已配置好所需依赖且运行正常。链接地址为：https:\u002F\u002Fcolab.research.google.com\u002Fgithub\u002Fneuroneural\u002Fbrainchop\u002Fblob\u002Fmaster\u002Fpy2tfjs\u002FMeshNet_Training_Example.ipynb","https:\u002F\u002Fgithub.com\u002Fneuroneural\u002Fbrainchop\u002Fissues\u002F70",{"id":161,"question_zh":162,"answer_zh":163,"source_url":164},22824,"上传的 MRI 图像归一化失败导致显示异常（数值未映射到 0-255）如何解决？","这是一个已知问题，部分来自 OpenNeuro 等来源的特定 MRI 文件（如 sub-01_ses-01_acq-MEMPRvNav_rec-RMS_T1w.nii.gz）可能无法正确归一化到 0-255 范围，导致复合显示损坏。虽然模型通常仍能运行并产生合理结果，但可视化会受影响。维护者已确认该 Bug 存在并在持续修复中，建议遇到此问题时检查是否有更新版本或尝试其他预处理步骤。","https:\u002F\u002Fgithub.com\u002Fneuroneural\u002Fbrainchop\u002Fissues\u002F22",{"id":166,"question_zh":167,"answer_zh":168,"source_url":169},22825,"Cortical Atlas 50 (failsafe) 模型运行失败如何处理？","该问题已在后续的代码提交中修复（Commit: e73a51f12ca50fe525b31a7ed8e1270d8095d66b）。如果您遇到此错误，请确保拉取了最新的代码库或等待网站端更新应用该修复补丁。","https:\u002F\u002Fgithub.com\u002Fneuroneural\u002Fbrainchop\u002Fissues\u002F23",[171,176,181,186,191,196,201,206,211,216,221,226,231,236,241,246,251,256,261,266],{"id":172,"version":173,"summary_zh":174,"released_at":175},136521,"v4.1.0","在 v4.1.0 中，新增了以下修复：\n\n- 修复 BrainChop 参数问题（[b3372b8](https:\u002F\u002Fgithub.com\u002Fneuroneural\u002Fbrainchop\u002Fcommit\u002Fb3372b8dc02115f7c62683b8b6bdabd0b6b570c8)）\n\n- 修复 104 高内存和快速模型参数问题（[e36437a](https:\u002F\u002Fgithub.com\u002Fneuroneural\u002Fbrainchop\u002Fcommit\u002Fe36437a1b18a0d36128026286f9d6e2434d11df1)）\n\n- 代码清理（例如 [1fc13a0](https:\u002F\u002Fgithub.com\u002Fneuroneural\u002Fbrainchop\u002Fcommit\u002F1fc13a044ed615c77fda0709da7b374cf4169025)）。","2024-06-02T22:56:26",{"id":177,"version":178,"summary_zh":179,"released_at":180},136522,"v4.0.0","# 发布说明\n\n## 版本 4.0.0\n\n我们非常高兴地欢迎 Chris Rorden (@neurolabusc) 和 Taylor Hanayik (@hanayik) 成为 brainchop 的新贡献者。特别感谢他们对 4.0 版本的主要贡献（[67396d2](https:\u002F\u002Fgithub.com\u002Fneuroneural\u002Fbrainchop\u002Fcommit\u002F67396d2651e0bd94c9dc877405cb0e2b0364d8c1)）。\n\n发布监督：@sergeyplis\n\n合并与配置：@Mmasoud1\n\n### 新特性\n- 集成了 [NiiVue](https:\u002F\u002Fgithub.com\u002Fniivue\u002Fniivue) 查看器，用于加载和可视化数据。支持拖放功能，可处理多种体素格式，包括 [NIfTI](https:\u002F\u002Fbrainder.org\u002F2012\u002F09\u002F23\u002Fthe-nifti-file-format\u002F)、[NRRD](http:\u002F\u002Fteem.sourceforge.net\u002Fnrrd\u002Fformat.html)、[MRtrix MIF](https:\u002F\u002Fmrtrix.readthedocs.io\u002Fen\u002Flatest\u002Fgetting_started\u002Fimage_data.html#mrtrix-image-formats)、[AFNI HEAD\u002FBRIK](https:\u002F\u002Fafni.nimh.nih.gov\u002Fpub\u002Fdist\u002Fdoc\u002Fprogram_help\u002FREADME.attributes.html)、[MGH\u002FMGZ](https:\u002F\u002Fsurfer.nmr.mgh.harvard.edu\u002Ffswiki\u002FFsTutorial\u002FMghFormat)、[ITK MHD](https:\u002F\u002Fitk.org\u002FWiki\u002FITK\u002FMetaIO\u002FDocumentation#Reading_a_Brick-of-Bytes_.28an_N-Dimensional_volume_in_a_single_file.29)、[ECAT7](https:\u002F\u002Fgithub.com\u002Fopenneuropet\u002FPET2BIDS\u002Ftree\u002F28aae3fab22309047d36d867c624cd629c921ca6\u002Fecat_validation\u002Fecat_info)，并提供有限的 DICOM 支持。\n  \n- 允许模型在主线程或 Web Worker 中运行。\n- 允许通过绘制来添加或擦除体素，从而编辑分类结果。\n- 可将当前场景（图像、分类结果、对比度、十字线位置、绘制内容）保存为 NiiVue 文档，可在任何地方打开。\n\n### 功能改进\n- 进行重构以提升性能。\n- 优化了预处理和后处理步骤。\n- 移除了 conform 函数的 Python 依赖（pyodide），并用原生 JavaScript 实现该功能。\n- 降低了静态内存占用。\n- 极大地减少了堆内存峰值使用量（对于低内存的 104 区域模型，减少幅度达 173 倍）。","2024-05-18T09:52:16",{"id":182,"version":183,"summary_zh":184,"released_at":185},136523,"v3.4.0","在 v3.4.0 中，新增了以下修复：\n\n- 顺序低内存模型的运行速度提升至2倍（[2e555f4](https:\u002F\u002Fgithub.com\u002Fneuroneural\u002Fbrainchop\u002Fcommit\u002F2e555f4b0d91ac85c7c2eb8b40d0ba8ae120249b)）","2024-05-12T03:19:40",{"id":187,"version":188,"summary_zh":189,"released_at":190},136524,"v3.3.0","在 v3.3.0 中，新增了以下修复：\n\n- 加速连通组件后处理（[89a01c9](https:\u002F\u002Fgithub.com\u002Fneuroneural\u002Fbrainchop\u002Fcommit\u002F89a01c9dbd3ee3087d195363b5763cb45849033c)）\n\n","2024-05-07T19:07:27",{"id":192,"version":193,"summary_zh":194,"released_at":195},136525,"v3.2.1","在 v3.2.1 中，新增了以下修复：\n- 修复缓冲区泄漏问题（[03d4019](https:\u002F\u002Fgithub.com\u002Fneuroneural\u002Fbrainchop\u002Fpull\u002F34\u002Fcommits\u002F03d4019cea8a356e3d5a91b5628d525dfa88963e)）\n- 修复归一化问题 #22（[b17d9c6](https:\u002F\u002Fgithub.com\u002Fneuroneural\u002Fbrainchop\u002Fcommit\u002Fb17d9c644a7438a6ae23f9498ddf6ba54c854a1f)）","2024-04-10T08:35:22",{"id":197,"version":198,"summary_zh":199,"released_at":200},136526,"v3.2.0","在 v3.2.0 中，新增了以下功能和修复：\n- 添加新的故障保护皮下模型，使用 preModel 进行裁剪 ([0fe66a3](https:\u002F\u002Fgithub.com\u002Fneuroneural\u002Fbrainchop\u002Fcommit\u002F0fe66a3ad8ca2dffb7ecb5983657cb4e17ed6915#diff-ec69380ca9a43bac83b16dc0fd4a425bc22decdb163f365748685eeb918605ac))\n- 启用生产模式 ([0fe66a3](https:\u002F\u002Fgithub.com\u002Fneuroneural\u002Fbrainchop\u002Fcommit\u002F0fe66a3ad8ca2dffb7ecb5983657cb4e17ed6915#diff-2023ef8640e47171ff04e2d34d456812da3975ba81cbafd552c40344846fbab6))\n- 为轻量级模型开启 enableCrop 功能 ([0fe66a3](https:\u002F\u002Fgithub.com\u002Fneuroneural\u002Fbrainchop\u002Fcommit\u002F0fe66a3ad8ca2dffb7ecb5983657cb4e17ed6915#diff-ec69380ca9a43bac83b16dc0fd4a425bc22decdb163f365748685eeb918605ac))","2024-03-30T20:17:09",{"id":202,"version":203,"summary_zh":204,"released_at":205},136527,"v3.1.0","在 v3.1.0 中，添加了以下修复：\r\n#### 修复\r\n\r\n\r\n- 修复了使用 f16 位纹理时的 3D 连通组件错误（[eff81f3](https:\u002F\u002Fgithub.com\u002Fneuroneural\u002Fbrainchop\u002Fpull\u002F31\u002Fcommits\u002Feff81f39bd348b5bac9d400c1008b6ae74ce7f72)）\r\n- 为皮层下模型关闭 preModel 裁剪功能（[71a1a49](https:\u002F\u002Fgithub.com\u002Fneuroneural\u002Fbrainchop\u002Fpull\u002F31\u002Fcommits\u002F71a1a4967ae732bad6dc9c24468c6f6ec8a427d3)）","2024-03-30T18:19:15",{"id":207,"version":208,"summary_zh":209,"released_at":210},136528,"v3.0.0","在 v3.0.0 版本中，我们新增了以下功能：\r\n#### 功能：\r\n\r\n\r\n- 添加更鲁棒的模型（[2b061eb](https:\u002F\u002Fgithub.com\u002Fneuroneural\u002Fbrainchop\u002Fcommit\u002F2b061ebca2cbe511ff4e208b68b6ca8b008e6433)）\r\n- 添加 PyTorch 到 TF.js 的转换示例（[7840eec](https:\u002F\u002Fgithub.com\u002Fneuroneural\u002Fbrainchop\u002Fcommit\u002F7840eec894b2cd602817e2ffdcba8c5ee6abb011)）","2024-03-28T11:08:55",{"id":212,"version":213,"summary_zh":214,"released_at":215},136529,"v2.4.0","在 v2.4.0 版本中，我们增加了更多选项：\r\n#### 功能特性：\r\n\r\n- 添加了 18 类模型（[f0d3153](https:\u002F\u002Fgithub.com\u002Fneuroneural\u002Fbrainchop\u002Fcommit\u002Ff0d3153a526e9a7f3424c894e4f0a28f4d0cfda1)）。","2024-03-27T02:04:42",{"id":217,"version":218,"summary_zh":219,"released_at":220},136530,"v2.3.0","在 v2.3.0 版本中，我们增加了更多选项，并修复了以下问题：\n\n#### 修复内容：\n\n- 修复内存问题（[baa236b](https:\u002F\u002Fgithub.com\u002Fneuroneural\u002Fbrainchop\u002Fcommit\u002Fbaa236b85b7d553c2bc90df41f412d2b99ca9c98)）\n\n- 添加适用于低内存资源的序列卷积选项（[1c4f657](https:\u002F\u002Fgithub.com\u002Fneuroneural\u002Fbrainchop\u002Fcommit\u002F1c4f65725522392a3254464d21f0cc02fdab01723)）","2024-03-19T17:28:55",{"id":222,"version":223,"summary_zh":224,"released_at":225},136531,"v2.2.0","In v2.2.0, we addressed the following minor fixes and adding a Colab example for DL training:\r\n\r\n #### Fixes:\r\n\r\n- Fix  .nii extension with mri_convert ([fb8cccb](https:\u002F\u002Fgithub.com\u002Fneuroneural\u002Fbrainchop\u002Fcommit\u002Ffb8cccb34eec88c987e7e94dd76b1a280f034bf1))  \r\n\r\n- Add Colab Example for MeshNet training  ([87a1ba9](https:\u002F\u002Fgithub.com\u002Fneuroneural\u002Fbrainchop\u002Fcommit\u002F87a1ba98cba92a041693402ef9169f28f5572bc3))\r\n\r\n- Fix unit testing links ([441f5be](https:\u002F\u002Fgithub.com\u002Fneuroneural\u002Fbrainchop\u002Fcommit\u002F441f5bec38e0b5558c66c96bd7b7a4da1e2658d1))","2024-01-20T17:07:13",{"id":227,"version":228,"summary_zh":229,"released_at":230},136532,"v2.1.0","In v2.1.0, we addressed the rest of JOSS reviewers reported bugs in addition to those addressed in v2.0.1.\r\n\r\n #### Fixes:\r\n\r\n- Adjusted padding to fix Papaya MRI viewers planes visibility for  screens of width > 1350  ([1c9f924](https:\u002F\u002Fgithub.com\u002Fneuroneural\u002Fbrainchop\u002Fcommit\u002F1c9f92411828db9d8168686a6c35f34474d11d07)).\r\n\r\n- Added two deeper Meshnet models for better brain extraction and masking accuracy ([0c6f90e](https:\u002F\u002Fgithub.com\u002Fneuroneural\u002Fbrainchop\u002Fcommit\u002F0c6f90e7ae8bc9e9d4905554aef28e1cd422ea0b)). \r\n\r\n- Fixed Mocha testing for  normalizeVolumeData function ([60c2240](https:\u002F\u002Fgithub.com\u002Fneuroneural\u002Fbrainchop\u002Fcommit\u002F60c224041825efdb16d1b6e9fa4a40132822f49a)).\r\n\r\n- Modified info and  warning messages text ([be8e0a8](https:\u002F\u002Fgithub.com\u002Fneuroneural\u002Fbrainchop\u002Fcommit\u002Fbe8e0a8c4ee4c1ef6ae48c69aedfd19da62722a8)).","2023-03-15T08:26:54",{"id":232,"version":233,"summary_zh":234,"released_at":235},136533,"v2.0.1"," #### Fixes:\r\n\r\n\r\n- ِِAdd scrollbar to the left mini forms container ([e951fbd](https:\u002F\u002Fgithub.com\u002Fneuroneural\u002Fbrainchop\u002Fcommit\u002Fe951fbd33d540fd131c32357482387d2275675bd)) (#18 )\r\n\r\n-  Fix models info  and warning  tooltips ([9f5cae7](https:\u002F\u002Fgithub.com\u002Fneuroneural\u002Fbrainchop\u002Fcommit\u002F9f5cae7e0e52dc8597906992ad3ad99974062ac0)) (#19 )","2023-03-08T01:13:31",{"id":237,"version":238,"summary_zh":239,"released_at":240},136534,"v2.0.0","With  v2.0.0, it can be said that brainchop in a relatively short time  exceeded the expectations as the first front-end neuroimaging tool for volumetric MRI processing. Many features have been added to brainchop functionality since first version to support data residency, 3D data rendering, preprocessing, volumetric segmentation, and postprocessing for the first time in the browser.   We hope that proof of concept will pave the road for next-generation neuroimaging applications in the browser. \r\n\r\n\r\n #### Features:\r\n\r\n- Add Mocha unit testing ([accac1f](https:\u002F\u002Fgithub.com\u002Fneuroneural\u002Fbrainchop\u002Fcommit\u002Faccac1f9cf6a7f26df187818733adbfb84c4f20a)).\r\n\r\n- Create draft-pdf yml file. ([a3f88e3](https:\u002F\u002Fgithub.com\u002Fneuroneural\u002Fbrainchop\u002Fcommit\u002Fa3f88e3dd3357859b74be8cbc4424fc8d77b06df)).\r\n\r\n- Add check browser resources option. ([b877464](https:\u002F\u002Fgithub.com\u002Fneuroneural\u002Fbrainchop\u002Fcommit\u002Fb877464f88341e4964116a6d4f7e0f2aebf985a6)).\r\n\r\n- Add list of verified H\u002FW and S\u002FW. ([b877464](https:\u002F\u002Fgithub.com\u002Fneuroneural\u002Fbrainchop\u002Fcommit\u002Fb877464f88341e4964116a6d4f7e0f2aebf985a6)).","2023-02-21T05:04:42",{"id":242,"version":243,"summary_zh":244,"released_at":245},136535,"v1.4.0"," #### Features:\r\n\r\n- Add 3D real-time rendering of the input Nifty file. ([44485e8](https:\u002F\u002Fgithub.com\u002Fneuroneural\u002Fbrainchop\u002Fcommit\u002F44485e8f168e3c4cdd16636d23159958db8722d9)).\r\n\r\n- Add 3D Image processing algorithms for input denoising and enhancing. ([44485e8](https:\u002F\u002Fgithub.com\u002Fneuroneural\u002Fbrainchop\u002Fcommit\u002F44485e8f168e3c4cdd16636d23159958db8722d9)).","2022-09-12T21:03:01",{"id":247,"version":248,"summary_zh":249,"released_at":250},136536,"v1.3.0"," #### Features:\r\n\r\n- Add three.js with WebGL backend to support T1 MRI rendering and filtering regions of interest in real time. ([a2c9842](https:\u002F\u002Fgithub.com\u002Fneuroneural\u002Fbrainchop\u002Fcommit\u002Fa2c9842ba01109253e37b1a95af87e22777e51c5)).\r\n\r\n- Create three.js dat.GUI checkbox list extension to support selecting multiple segmented region to 3D visualization. ([f76eee2](https:\u002F\u002Fgithub.com\u002Fneuroneural\u002Fbrainchop\u002Fcommit\u002Ff76eee2fb11b4a16942d9a43a02a682aaebd9a09)).\r\n\r\n- Add bar chart to plot Meshnet segmentation output volumes. ([a2c9842](https:\u002F\u002Fgithub.com\u002Fneuroneural\u002Fbrainchop\u002Fcommit\u002Fa2c9842ba01109253e37b1a95af87e22777e51c5)).\r\n\r\n- Adjust layout to add new options.([a2c9842](https:\u002F\u002Fgithub.com\u002Fneuroneural\u002Fbrainchop\u002Fcommit\u002Fa2c9842ba01109253e37b1a95af87e22777e51c5)) ","2022-08-04T20:10:17",{"id":252,"version":253,"summary_zh":254,"released_at":255},136537,"v1.2.1"," #### Fixes:\r\n\r\n\r\n- Fix environment parameters ([2cfc9d7](https:\u002F\u002Fgithub.com\u002Fneuroneural\u002Fbrainchop\u002Fcommit\u002F2cfc9d7005ec64b77cfafa777948fc0faa29b513)).","2022-08-01T03:24:33",{"id":257,"version":258,"summary_zh":259,"released_at":260},136538,"v1.2.0"," #### Features:\r\n\r\n\r\n- Add 50 segmentation model to segment MRI cortical regions. Each cortical structure region is marked with a unique color\u002Flabel  compatible with FreeSurfer ([8375a08](https:\u002F\u002Fgithub.com\u002Fneuroneural\u002Fbrainchop\u002Fcommit\u002F8375a08c514a7f75b1f263fa8007e8fcb7f57e96)).\r\n- Add 104 segmentation model to segment MRI cortical and sub-cortical structures. Each structure region is marked with a unique color\u002Flabel  compatible with FreeSurfer ([5c0e388](https:\u002F\u002Fgithub.com\u002Fneuroneural\u002Fbrainchop\u002Fcommit\u002F5c0e388ecccdcac93497996db160e6ac861bc7e0)).\r\n- Add MRI tissue cropping pipeline to speedup the inference and lowering the memory use  ([8375a08](https:\u002F\u002Fgithub.com\u002Fneuroneural\u002Fbrainchop\u002Fcommit\u002F8375a08c514a7f75b1f263fa8007e8fcb7f57e96)).\r\n- Add  new  Buffer  class supports 'uint8'|'int8'|'uint16'|'int16'| 'float16'  to minimize memory overhead ([7ce947b](https:\u002F\u002Fgithub.com\u002Fneuroneural\u002Fbrainchop\u002Fcommit\u002F7ce947bee04d83d5e0566aecbcf7a2405b6160cf)) and  ([5c0e388](https:\u002F\u002Fgithub.com\u002Fneuroneural\u002Fbrainchop\u002Fcommit\u002F5c0e388ecccdcac93497996db160e6ac861bc7e0)).\r\n- Fix memory leak ([3f379ea](https:\u002F\u002Fgithub.com\u002Fneuroneural\u002Fbrainchop\u002Fcommit\u002F3f379ea3137060b26efc9f06d125527245b924ac)).\r\n\r\n","2022-07-19T23:30:34",{"id":262,"version":263,"summary_zh":264,"released_at":265},136539,"v1.1.0"," #### Features:\r\n\r\n- Add  mri_convert,   a pre-processing converter to support reshaping\u002Fnormalizing\u002Fresampling MRI raw input data  ([7b6ff1c](https:\u002F\u002Fgithub.com\u002Fneuroneural\u002Fbrainchop\u002Fcommit\u002F7b6ff1c282349b47b439ebf11e05e8f6c49b5715))","2022-06-16T05:49:38",{"id":267,"version":268,"summary_zh":269,"released_at":270},136540,"v1.0.0","We are excited to announce the first release of brainchop. This brings automatic 3D MRI segmentation capability to neuroimaging. You can use the online front-end brainchop.org or use it in the offline mode. Offline setup instructions are available on brainchop Wiki.\r\n\r\n#### This package:\r\n\r\n- Supports inference on 3D volumetric MRI data with WebGL.\r\n- Supports full brain volume inference and subcube inference.\r\n- Supports post-processing and noise removal.\r\n- Supports saving resulting MRI labels in the Nifti format.\r\n- Supports pre-processing and normalization\r\n- Supports importing compatible pre-trained custom models in tfjs.\r\n\r\n\r\nWe welcome your feedback to help shape our priorities for brainchop. We also welcome contributors familiar with tfjs and neuroimaging applications who are interested in getting involved in expanding brainchop model zoo.\r\n\r\nLooking forward to your contribution!","2022-04-27T03:26:41"]