[{"data":1,"prerenderedAt":-1},["ShallowReactive",2],{"similar-the-super-toys--glimpse-android":3,"tool-the-super-toys--glimpse-android":64},[4,17,27,35,43,56],{"id":5,"name":6,"github_repo":7,"description_zh":8,"stars":9,"difficulty_score":10,"last_commit_at":11,"category_tags":12,"status":16},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。",162132,3,"2026-04-05T11:01:52",[13,14,15],"开发框架","图像","Agent","ready",{"id":18,"name":19,"github_repo":20,"description_zh":21,"stars":22,"difficulty_score":23,"last_commit_at":24,"category_tags":25,"status":16},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",138956,2,"2026-04-05T11:33:21",[13,15,26],"语言模型",{"id":28,"name":29,"github_repo":30,"description_zh":31,"stars":32,"difficulty_score":23,"last_commit_at":33,"category_tags":34,"status":16},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。",107662,"2026-04-03T11:11:01",[13,14,15],{"id":36,"name":37,"github_repo":38,"description_zh":39,"stars":40,"difficulty_score":23,"last_commit_at":41,"category_tags":42,"status":16},3704,"NextChat","ChatGPTNextWeb\u002FNextChat","NextChat 是一款轻量且极速的 AI 助手，旨在为用户提供流畅、跨平台的大模型交互体验。它完美解决了用户在多设备间切换时难以保持对话连续性，以及面对众多 AI 模型不知如何统一管理的痛点。无论是日常办公、学习辅助还是创意激发，NextChat 都能让用户随时随地通过网页、iOS、Android、Windows、MacOS 或 Linux 端无缝接入智能服务。\n\n这款工具非常适合普通用户、学生、职场人士以及需要私有化部署的企业团队使用。对于开发者而言，它也提供了便捷的自托管方案，支持一键部署到 Vercel 或 Zeabur 等平台。\n\nNextChat 的核心亮点在于其广泛的模型兼容性，原生支持 Claude、DeepSeek、GPT-4 及 Gemini Pro 等主流大模型，让用户在一个界面即可自由切换不同 AI 能力。此外，它还率先支持 MCP（Model Context Protocol）协议，增强了上下文处理能力。针对企业用户，NextChat 提供专业版解决方案，具备品牌定制、细粒度权限控制、内部知识库整合及安全审计等功能，满足公司对数据隐私和个性化管理的高标准要求。",87618,"2026-04-05T07:20:52",[13,26],{"id":44,"name":45,"github_repo":46,"description_zh":47,"stars":48,"difficulty_score":23,"last_commit_at":49,"category_tags":50,"status":16},2268,"ML-For-Beginners","microsoft\u002FML-For-Beginners","ML-For-Beginners 是由微软推出的一套系统化机器学习入门课程，旨在帮助零基础用户轻松掌握经典机器学习知识。这套课程将学习路径规划为 12 周，包含 26 节精炼课程和 52 道配套测验，内容涵盖从基础概念到实际应用的完整流程，有效解决了初学者面对庞大知识体系时无从下手、缺乏结构化指导的痛点。\n\n无论是希望转型的开发者、需要补充算法背景的研究人员，还是对人工智能充满好奇的普通爱好者，都能从中受益。课程不仅提供了清晰的理论讲解，还强调动手实践，让用户在循序渐进中建立扎实的技能基础。其独特的亮点在于强大的多语言支持，通过自动化机制提供了包括简体中文在内的 50 多种语言版本，极大地降低了全球不同背景用户的学习门槛。此外，项目采用开源协作模式，社区活跃且内容持续更新，确保学习者能获取前沿且准确的技术资讯。如果你正寻找一条清晰、友好且专业的机器学习入门之路，ML-For-Beginners 将是理想的起点。",84991,"2026-04-05T10:45:23",[14,51,52,53,15,54,26,13,55],"数据工具","视频","插件","其他","音频",{"id":57,"name":58,"github_repo":59,"description_zh":60,"stars":61,"difficulty_score":10,"last_commit_at":62,"category_tags":63,"status":16},3128,"ragflow","infiniflow\u002Fragflow","RAGFlow 是一款领先的开源检索增强生成（RAG）引擎，旨在为大语言模型构建更精准、可靠的上下文层。它巧妙地将前沿的 RAG 技术与智能体（Agent）能力相结合，不仅支持从各类文档中高效提取知识，还能让模型基于这些知识进行逻辑推理和任务执行。\n\n在大模型应用中，幻觉问题和知识滞后是常见痛点。RAGFlow 通过深度解析复杂文档结构（如表格、图表及混合排版），显著提升了信息检索的准确度，从而有效减少模型“胡编乱造”的现象，确保回答既有据可依又具备时效性。其内置的智能体机制更进一步，使系统不仅能回答问题，还能自主规划步骤解决复杂问题。\n\n这款工具特别适合开发者、企业技术团队以及 AI 研究人员使用。无论是希望快速搭建私有知识库问答系统，还是致力于探索大模型在垂直领域落地的创新者，都能从中受益。RAGFlow 提供了可视化的工作流编排界面和灵活的 API 接口，既降低了非算法背景用户的上手门槛，也满足了专业开发者对系统深度定制的需求。作为基于 Apache 2.0 协议开源的项目，它正成为连接通用大模型与行业专有知识之间的重要桥梁。",77062,"2026-04-04T04:44:48",[15,14,13,26,54],{"id":65,"github_repo":66,"name":67,"description_en":68,"description_zh":69,"ai_summary_zh":69,"readme_en":70,"readme_zh":71,"quickstart_zh":72,"use_case_zh":73,"hero_image_url":74,"owner_login":75,"owner_name":76,"owner_avatar_url":77,"owner_bio":78,"owner_company":79,"owner_location":79,"owner_email":79,"owner_twitter":79,"owner_website":79,"owner_url":80,"languages":81,"stars":90,"forks":91,"last_commit_at":92,"license":93,"difficulty_score":23,"env_os":94,"env_gpu":95,"env_ram":96,"env_deps":97,"category_tags":104,"github_topics":105,"view_count":10,"oss_zip_url":79,"oss_zip_packed_at":79,"status":16,"created_at":109,"updated_at":110,"faqs":111,"releases":141},730,"the-super-toys\u002Fglimpse-android","glimpse-android","A content-aware cropping library for Android","glimpse-android 是一款专为 Android 平台打造的图片智能裁剪开源库。在移动端开发中，传统的图片裁剪往往采用简单的中心对齐方式，容易导致重要内容被误切，影响视觉效果。glimpse-android 通过引入人工智能技术，能够像人眼一样智能识别图片中的视觉焦点，从而计算出最佳的裁剪区域，确保展示的画面始终聚焦于核心内容。\n\nglimpse-android 特别适合 Android 应用程序开发者使用，尤其是那些希望提升图片展示质量、优化用户第一印象的团队。其独特的技术亮点在于底层集成了 TensorFlow Lite 模型，实现了真正的“内容感知”裁剪。此外，它还提供了对 Glide 和 Coil 等主流图片加载框架的扩展支持，开发者只需添加少量代码即可轻松启用智能裁剪功能，无需关心复杂的坐标计算逻辑。虽然目前尚未支持 Picasso 和 Fresco，但其灵活的 API 设计为未来扩展留下了空间。对于追求高品质 UI 体验的开发者来说，glimpse-android 是一个值得尝试的高效解决方案。","# Glimpse\n\nA content-aware cropping library for Android.\n\nGive the right first impression with just a glimpse! Instead of center cropping images blindly, leverage Glimpse's eye to catch the right spot.\n\n![image](https:\u002F\u002Foss.gittoolsai.com\u002Fimages\u002Fthe-super-toys_glimpse-android_readme_fe56ac8a730e.png)\n\n## Setup\nAdd to top level *gradle.build* file\n```gradle\nallprojects {\n    repositories {\n        maven { url \"https:\u002F\u002Fjitpack.io\" }\n    }\n}\n```\n\nAdd to app module *gradle.build* file\nTensorFlow lite recommends most developers omit the x86, x86_64, and arm32 ABIs. This can be achieved with the following Gradle configuration, which specifically includes only armeabi-v7a and arm64-v8a, which should cover most modern Android devices.\n```gradle\n\nandroid {\n    aaptOptions {\n        noCompress \"tflite\"\n        noCompress \"lite\"\n    }\n    defaultConfig {\n        ndk {\n            abiFilters 'armeabi-v7a', 'arm64-v8a'\n        }\n    }\n}\n\ndependencies {\n    implementation 'com.github.the-super-toys.glimpse-android:glimpse-core:0.0.5'\n    \n    \u002F\u002Fonly required for glide extensions\n    implementation 'com.github.the-super-toys.glimpse-android:glimpse-glide:0.0.5'\n    \n    \u002F\u002Fonly required for coil extensions\n    implementation 'com.github.the-super-toys.glimpse-android:glimpse-coil:0.0.5'\n    \n    implementation 'org.tensorflow:tensorflow-lite:0.0.0-nightly'\n}\n```\n\n## Usage\n\n### Init Glimpse in your Android app\n\n```kotlin\nclass YourApp : Application() {\n    override fun onCreate() {\n        super.onCreate()\n        Glimpse.init(this)\n    }\n}\n```\n\n### Use Glimpse without extensions\nFirst compute image's focal points by calling `Bitmap.findCenter` extension function.\n\n```kotlin\nval original = (imageView.drawable as BitmapDrawable).bitmap\nval (x, y) = original.findCenter()\n```\n\nThen supply those focal points alongside with the width and height values of the target `ImageView` to `Bitmap.crop` extension function.\n\n```kotlin\nval cropped = original.crop(x, y, imageView.layoutParams.width, imageView.layoutParams.height)\nimageView.setImageBitmap(cropped)\n```\n\n### Use Glimpse with Glide extension \nIf you are using [Glide](https:\u002F\u002Fgithub.com\u002Fbumptech\u002Fglide) for image loading you can just add `GlimpseTransformation` to `GlideRequest` builder:\n\n \n```kotlin\nGlideApp.with(context)\n                .load(url)\n                .diskCacheStrategy(DiskCacheStrategy.RESOURCE)\n                .transform(GlimpseTransformation())\n                .into(imageView)\n```\n\nIt is recommended to set `diskCacheStrategy(DiskCacheStrategy.RESOURCE)` to cache the cropped bitmap by Glimpse, otherwise focal points will be calculated every time the image is displayed.\n\n\n### Use Glimpse with Coil extension \nIf you are using [Coil](https:\u002F\u002Fgithub.com\u002Fcoil-kt\u002Fcoil) for image loading you can just add `GlimpseTransformation` to Coil's `LoadRequestBuilder` builder:\n\n \n```kotlin\nimageView.load(url) {\n    crossfade(true)\n    placeholder(R.drawable.image)\n    transformations(GlimpseTransformation())\n}\n\n\n```\n\n### What about other image loaders such as Picasso and Fresco?\n\nWe tried to ship Glimpse with both Picasso and Fresco extensions but we were not able to find the right api.\n\nWe opened an [issue on Picasso repository](https:\u002F\u002Fgithub.com\u002Fsquare\u002Fpicasso\u002Fissues\u002F2067) and a [labeled Fresco question on Stack Overflow](https:\u002F\u002Fstackoverflow.com\u002Fquestions\u002F54773198\u002Ffresco-how-to-use-scaletypes-focuscrop-based-on-bitmap-content)  asking for guidance, but sadly we did not find much support. If you know how to assemble Glimpse cropping functionality to either lib without disrupting their original pipeline, please open an issue to review together the proposal. We really want to offer support for all the popular loading image libraries!  \n  \n### Java callers\nIf you're using Java you can take a look at this [file](https:\u002F\u002Fgithub.com\u002Fthe-super-toys\u002Fglimpse-android\u002Fblob\u002Fmaster\u002Fsample-app\u002Fsrc\u002Fmain\u002Fjava\u002Fglimpse\u002Fsample\u002FTestingApiFromJava.java) to see how Glimpse api looks for Java callers. \n\n## Sample app\n`:sample-app` module showcase Glimpse usage by making use of Glide extensions. The application is also published on [Google Play](https:\u002F\u002Fplay.google.com\u002Fstore\u002Fapps\u002Fdetails?id=glimpse.sample). \n\n\n## Is Glimpse ready for production? \n\nActually, it depends. Ideally you should not use Glimpse to crop the same image over and over. Even if you use Glide extension which caches the output transformation, the underlying calculation \nwill be performed in every android client. Thus, if possible, send to your cloud solution the x and y focal points of the calculated crop alongside the image when user uploads content.\n\nWe timed the crop calculation on an OnePlus 6 and in average it takes 30 milliseconds. Depending on the device and your specific use case Glimpse may or may not be ready for production. In any case we encourage you to provide feedback to help us make Glimpse a mature library.\n\n\n## How much increase APK's size?\n\nIf you're already using [TensorFlow lite](https:\u002F\u002Fwww.tensorflow.org\u002Flite) in your app, adding Glimpse costs only the size of the model, which is 148 KB. But chances are that you are not using it, so you have to take into account also the size of [TensorFlow lite](https:\u002F\u002Fwww.tensorflow.org\u002Flite\u002Foverview#tensorflow_lite_highlights), which is about 300 KB.\n\n\n## Let's go DEEPER (The Deep Learning Model)\n\nIf you want to know how Glimpse works in a more deeper lever, take a look at [this repository](https:\u002F\u002Fgithub.com\u002Fthe-super-toys\u002Fglimpse-models), where you can see the architecture and implementation of the Deep Learning model used by Glimpse. We explain how everything works in a much more technical level, and with all details, so you can reproduce it on your own, and improve it if you want.  \n\nBut if you do, don't hesitate to submit a Pull Request to that repository so we can keep improving Glimpse!!\n","# Glimpse\n\n一个用于 Android 的内容感知裁剪库。\n\n只需一瞥，即可呈现完美的第一印象！与其盲目地居中裁剪图片，不如利用 Glimpse 的“眼睛”捕捉到正确的焦点位置。\n\n![image](https:\u002F\u002Foss.gittoolsai.com\u002Fimages\u002Fthe-super-toys_glimpse-android_readme_fe56ac8a730e.png)\n\n## 配置\n添加到顶层 *build.gradle* 文件\n```gradle\nallprojects {\n    repositories {\n        maven { url \"https:\u002F\u002Fjitpack.io\" }\n    }\n}\n```\n\n添加到应用模块 *build.gradle* 文件\nTensorFlow Lite (轻量级 TensorFlow) 建议大多数开发者省略 x86、x86_64 和 arm32 ABI (应用程序二进制接口)。这可以通过以下 Gradle 配置实现，该配置仅包含 armeabi-v7a 和 arm64-v8a，应能覆盖大多数现代 Android 设备。\n```gradle\n\nandroid {\n    aaptOptions {\n        noCompress \"tflite\"\n        noCompress \"lite\"\n    }\n    defaultConfig {\n        ndk {\n            abiFilters 'armeabi-v7a', 'arm64-v8a'\n        }\n    }\n}\n\ndependencies {\n    implementation 'com.github.the-super-toys.glimpse-android:glimpse-core:0.0.5'\n    \n    \u002F\u002Fonly required for glide extensions\n    implementation 'com.github.the-super-toys.glimpse-android:glimpse-glide:0.0.5'\n    \n    \u002F\u002Fonly required for coil extensions\n    implementation 'com.github.the-super-toys.glimpse-android:glimpse-coil:0.0.5'\n    \n    implementation 'org.tensorflow:tensorflow-lite:0.0.0-nightly'\n}\n```\n\n## 使用\n\n### 在您的 Android 应用中初始化 Glimpse\n\n```kotlin\nclass YourApp : Application() {\n    override fun onCreate() {\n        super.onCreate()\n        Glimpse.init(this)\n    }\n}\n```\n\n### 不使用扩展功能使用 Glimpse\n首先通过调用 `Bitmap.findCenter` 扩展函数来计算图像的焦点。\n\n```kotlin\nval original = (imageView.drawable as BitmapDrawable).bitmap\nval (x, y) = original.findCenter()\n```\n\n然后将这些焦点连同目标 `ImageView` 的宽度和高度值提供给 `Bitmap.crop` 扩展函数。\n\n```kotlin\nval cropped = original.crop(x, y, imageView.layoutParams.width, imageView.layoutParams.height)\nimageView.setImageBitmap(cropped)\n```\n\n### 配合 Glide 扩展使用 Glimpse \n如果您使用 [Glide](https:\u002F\u002Fgithub.com\u002Fbumptech\u002Fglide) 进行图像加载，只需将 `GlimpseTransformation` 添加到 `GlideRequest` 构建器中：\n\n \n```kotlin\nGlideApp.with(context)\n                .load(url)\n                .diskCacheStrategy(DiskCacheStrategy.RESOURCE)\n                .transform(GlimpseTransformation())\n                .into(imageView)\n```\n\n建议设置 `diskCacheStrategy(DiskCacheStrategy.RESOURCE)` 以缓存由 Glimpse 裁剪后的位图，否则每次显示图像时都会重新计算焦点。\n\n\n### 配合 Coil 扩展使用 Glimpse \n如果您使用 [Coil](https:\u002F\u002Fgithub.com\u002Fcoil-kt\u002Fcoil) 进行图像加载，只需将 `GlimpseTransformation` 添加到 Coil 的 `LoadRequestBuilder` 构建器中：\n\n \n```kotlin\nimageView.load(url) {\n    crossfade(true)\n    placeholder(R.drawable.image)\n    transformations(GlimpseTransformation())\n}\n\n\n```\n\n### 其他图像加载库（如 Picasso 和 Fresco）怎么办？\n\n我们曾尝试为 Glimpse 提供 Picasso 和 Fresco 的扩展，但未能找到合适的 API (应用程序编程接口)。\n\n我们在 [Picasso 仓库上提交了 issue](https:\u002F\u002Fgithub.com\u002Fsquare\u002Fpicasso\u002Fissues\u002F2067)，并在 [Stack Overflow 上标记了关于 Fresco 的问题](https:\u002F\u002Fstackoverflow.com\u002Fquestions\u002F54773198\u002Ffresco-how-to-use-scaletypes-focuscrop-based-on-bitmap-content) 寻求指导，但遗憾的是我们没有得到太多支持。如果您知道如何将 Glimpse 的裁剪功能集成到任一库中而不破坏其原有流程，请提交 issue 以便我们一起审查提案。我们非常希望能为所有流行的图像加载库提供支持！  \n  \n### Java 调用者\n如果您使用 Java，可以查看此 [文件](https:\u002F\u002Fgithub.com\u002Fthe-super-toys\u002Fglimpse-android\u002Fblob\u002Fmaster\u002Fsample-app\u002Fsrc\u002Fmain\u002Fjava\u002Fglimpse\u002Fsample\u002FTestingApiFromJava.java)，了解 Glimpse API 对 Java 调用者的样子。 \n\n## 示例应用\n`:sample-app` 模块通过使用 Glide 扩展展示了 Glimpse 的使用方式。该应用也已发布在 [Google Play](https:\u002F\u002Fplay.google.com\u002Fstore\u002Fapps\u002Fdetails?id=glimpse.sample) 上。 \n\n\n## Glimpse 是否已准备好投入生产？ \n\n实际上，这取决于具体情况。理想情况下，您不应反复使用 Glimpse 裁剪同一张图片。即使您使用了会缓存输出转换的 Glide 扩展，底层计算仍会在每个 Android 客户端执行。因此，如果可能，当用户上传内容时，请将计算出的裁剪点的 x 和 y 坐标连同图像一起发送到您的云端解决方案。\n\n我们在 OnePlus 6 上测试了裁剪计算的时间，平均耗时 30 毫秒。根据设备和您的具体用例，Glimpse 可能适合也可能不适合投入生产。无论如何，我们鼓励您提供反馈，帮助我们使 Glimpse 成为成熟的库。\n\n\n## APK 体积会增加多少？\n\n如果您的应用中已经使用了 [TensorFlow Lite (轻量级 TensorFlow)](https:\u002F\u002Fwww.tensorflow.org\u002Flite)，添加 Glimpse 仅需增加模型的大小，即 148 KB。但很有可能您并未使用它，因此您还需要考虑 [TensorFlow Lite](https:\u002F\u002Fwww.tensorflow.org\u002Flite\u002Foverview#tensorflow_lite_highlights) 本身的大小，约为 300 KB。\n\n\n## 让我们深入探讨（深度学习模型）\n\n如果您想以更深的层次了解 Glimpse 的工作原理，请查看 [此仓库](https:\u002F\u002Fgithub.com\u002Fthe-super-toys\u002Fglimpse-models)，您可以看到 Glimpse 所使用的深度学习模型的架构和实现。我们以更技术化的层面解释了所有内容及其细节，以便您可以自行复现，并视情况改进它。  \n\n但如果您这样做了，请不要犹豫向该仓库提交 Pull Request (拉取请求)，以便我们不断改进 Glimpse！！","# glimpse-android 快速上手指南\n\n`glimpse-android` 是一个面向 Android 的内容感知裁剪库。它利用深度学习模型智能识别图像焦点，替代传统的盲目中心裁剪，为应用提供更优质的视觉体验。\n\n## 环境准备\n\n*   **开发工具**: Android Studio\n*   **构建系统**: Gradle\n*   **语言**: Kotlin 或 Java\n*   **NDK**: 需支持 armeabi-v7a 和 arm64-v8a 架构（TensorFlow Lite 推荐配置）\n\n## 安装步骤\n\n### 1. 配置仓库地址\n在项目的根目录 `build.gradle` 文件中添加 JitPack 仓库：\n\n```gradle\nallprojects {\n    repositories {\n        maven { url \"https:\u002F\u002Fjitpack.io\" }\n    }\n}\n```\n\n### 2. 配置模块依赖\n在应用模块的 `build.gradle` 文件中添加以下配置。建议仅保留主流 ARM 架构以减小体积。\n\n```gradle\nandroid {\n    aaptOptions {\n        noCompress \"tflite\"\n        noCompress \"lite\"\n    }\n    defaultConfig {\n        ndk {\n            abiFilters 'armeabi-v7a', 'arm64-v8a'\n        }\n    }\n}\n\ndependencies {\n    implementation 'com.github.the-super-toys.glimpse-android:glimpse-core:0.0.5'\n    \n    \u002F\u002F 如果使用 Glide 加载图片（可选）\n    implementation 'com.github.the-super-toys.glimpse-android:glimpse-glide:0.0.5'\n    \n    \u002F\u002F 如果使用 Coil 加载图片（可选）\n    implementation 'com.github.the-super-toys.glimpse-android:glimpse-coil:0.0.5'\n    \n    implementation 'org.tensorflow:tensorflow-lite:0.0.0-nightly'\n}\n```\n\n## 基本使用\n\n### 1. 初始化库\n在您的 `Application` 类中调用初始化方法：\n\n```kotlin\nclass YourApp : Application() {\n    override fun onCreate() {\n        super.onCreate()\n        Glimpse.init(this)\n    }\n}\n```\n\n### 2. 基础裁剪示例\n通过扩展函数计算焦点并进行裁剪：\n\n```kotlin\nval original = (imageView.drawable as BitmapDrawable).bitmap\nval (x, y) = original.findCenter()\n\nval cropped = original.crop(x, y, imageView.layoutParams.width, imageView.layoutParams.height)\nimageView.setImageBitmap(cropped)\n```\n\n### 3. 集成 Glide 加载器\n若使用 Glide 加载图片，可直接添加变换器并开启资源缓存策略：\n\n```kotlin\nGlideApp.with(context)\n                .load(url)\n                .diskCacheStrategy(DiskCacheStrategy.RESOURCE)\n                .transform(GlimpseTransformation())\n                .into(imageView)\n```\n\n> **提示**：设置 `diskCacheStrategy(DiskCacheStrategy.RESOURCE)` 可避免每次显示时重复计算焦点，提升性能。","某电商 APP 开发团队正在优化商品列表页的图片展示效果，需要处理大量用户上传的竖图与横图混排情况。\n\n### 没有 glimpse-android 时\n- 传统中心裁剪常把商品主体切掉一半，导致用户第一眼看不到重点。\n- 开发者需手动编写复杂的逻辑来识别图片内容，代码维护成本极高。\n- 不同尺寸的图片在网格布局中显得参差不齐，严重影响整体美观度。\n- 无法缓存智能裁剪结果，每次加载都重复计算，消耗设备性能且增加流量。\n\n### 使用 glimpse-android 后\n- glimpse-android 自动识别画面焦点，确保商品主体始终完整显示在视口内。\n- 集成 Glide 或 Coil 扩展仅需添加 Transformation，无需额外开发图像分析逻辑。\n- 配合资源缓存策略，避免重复计算，显著提升列表滑动流畅度。\n- 统一了所有商品图的视觉比例，让店铺页面看起来更加专业整洁。\n\nglimpse-android 通过 AI 智能聚焦解决了自适应布局下的主体丢失问题，极大提升了移动端图片展示的质感。","https:\u002F\u002Foss.gittoolsai.com\u002Fimages\u002Fthe-super-toys_glimpse-android_e0f2725a.png","the-super-toys","The Super Toys","https:\u002F\u002Foss.gittoolsai.com\u002Favatars\u002Fthe-super-toys_17663fb4.png","A Delusional Team",null,"https:\u002F\u002Fgithub.com\u002Fthe-super-toys",[82,86],{"name":83,"color":84,"percentage":85},"Kotlin","#A97BFF",96.9,{"name":87,"color":88,"percentage":89},"Java","#b07219",3.1,590,30,"2026-03-14T03:19:19","Apache-2.0","Android","无需桌面 GPU，基于移动端 NPU\u002FCPU，无 CUDA 要求","未说明",{"notes":98,"python":96,"dependencies":99},"这是一个 Android 内容感知裁剪库，非桌面端工具。依赖 TensorFlow Lite 模型（约 300KB）。在 OnePlus 6 上测试平均裁剪耗时 30ms。建议将焦点坐标上传至云端缓存以减少客户端计算压力。支持 Glide 和 Coil 扩展。",[100,101,102,103],"org.tensorflow:tensorflow-lite:0.0.0-nightly","com.github.the-super-toys.glimpse-android:glimpse-core:0.0.5","com.github.the-super-toys.glimpse-android:glimpse-glide:0.0.5","com.github.the-super-toys.glimpse-android:glimpse-coil:0.0.5",[13,14],[106,107,108],"machine-learning","android","crop-image","2026-03-27T02:49:30.150509","2026-04-06T07:14:22.792355",[112,117,121,126,131,136],{"id":113,"question_zh":114,"answer_zh":115,"source_url":116},3091,"调用 GlimpseTransformation 时报错，初始化语法有什么要求？","在初始化并调用 `GlimpseTransformation` 之前，必须先添加 `new` 关键字。例如使用 `new GlimpseTransformation()` 进行实例化。","https:\u002F\u002Fgithub.com\u002Fthe-super-toys\u002Fglimpse-android\u002Fissues\u002F16",{"id":118,"question_zh":119,"answer_zh":120,"source_url":116},3092,"集成库后仍无法工作，需要补充哪些 Gradle 依赖？","需要在项目的 Gradle 文件中显式添加 TensorFlow 相关依赖（如 `org.tensorflow:tensorflow-lite`），否则库无法正常运行。",{"id":122,"question_zh":123,"answer_zh":124,"source_url":125},3093,"是否支持将 minSdkVersion 设置为 15？","官方不再支持该版本。但可以通过在应用层 build.gradle 中使用 `tools:overrideLibrary` 来覆盖库的 minSdkVersion，从而实现兼容。","https:\u002F\u002Fgithub.com\u002Fthe-super-toys\u002Fglimpse-android\u002Fissues\u002F11",{"id":127,"question_zh":128,"answer_zh":129,"source_url":130},3094,"引入库后 APK 体积异常增大（超过 3MB）如何处理？","项目目前处于非活跃维护状态。体积增加主要源于 TensorFlow Lite 库本身。建议检查依赖配置，若无法优化需自行寻找解决方案或提交 PR。","https:\u002F\u002Fgithub.com\u002Fthe-super-toys\u002Fglimpse-android\u002Fissues\u002F13",{"id":132,"question_zh":133,"answer_zh":134,"source_url":135},3095,"运行时报错 `ArrayIndexOutOfBoundsException` 提示 Tensor 和像素数组形状不兼容怎么办？","这表示输入图像尺寸与张量期望形状不符。请检查传入函数的参数值，并提供设备型号、Android 版本及屏幕分辨率等信息以便排查。","https:\u002F\u002Fgithub.com\u002Fthe-super-toys\u002Fglimpse-android\u002Fissues\u002F7",{"id":137,"question_zh":138,"answer_zh":139,"source_url":140},3096,"使用 `debugHeatMap` 方法崩溃，提示 `y must be \u003C bitmap.height()` 如何解决？","这是调试方法，崩溃通常因 Bitmap 维度参数超出范围。请确认传入的 Bitmap 高度和坐标值是否正确，并提供具体复现信息以便修复。","https:\u002F\u002Fgithub.com\u002Fthe-super-toys\u002Fglimpse-android\u002Fissues\u002F1",[142],{"id":143,"version":144,"summary_zh":145,"released_at":146},102634,"0.0.5","- Added support for Coil, kudos to @aniketbhoite \r\n- Updated the readme to explain how to import tf-lite with a very small memory footprint @aniketbhoite ","2020-02-02T11:04:54"]