{"id":13436725,"url":"https://github.com/AIDajiangtang/Segment-Anything-CPP","last_synced_at":"2025-03-18T21:30:59.653Z","repository":{"id":180880697,"uuid":"665848103","full_name":"AIDajiangtang/Segment-Anything-CPP","owner":"AIDajiangtang","description":"segment anything（SAM） for CPP Inference","archived":false,"fork":false,"pushed_at":"2024-06-11T07:35:38.000Z","size":1554,"stargazers_count":29,"open_issues_count":9,"forks_count":2,"subscribers_count":1,"default_branch":"main","last_synced_at":"2024-10-27T20:20:56.857Z","etag":null,"topics":["deep-learning","segment-anything"],"latest_commit_sha":null,"homepage":"","language":"C++","has_issues":true,"has_wiki":null,"has_pages":null,"mirror_url":null,"source_name":null,"license":null,"status":null,"scm":"git","pull_requests_enabled":true,"icon_url":"https://github.com/AIDajiangtang.png","metadata":{"files":{"readme":"README.md","changelog":null,"contributing":null,"funding":null,"license":null,"code_of_conduct":null,"threat_model":null,"audit":null,"citation":null,"codeowners":null,"security":null,"support":null,"governance":null,"roadmap":null,"authors":null,"dei":null,"publiccode":null,"codemeta":null}},"created_at":"2023-07-13T06:24:44.000Z","updated_at":"2024-09-13T03:35:10.000Z","dependencies_parsed_at":"2024-10-27T19:14:18.111Z","dependency_job_id":"9cdce415-761d-4b3f-883c-1298a6f18ca0","html_url":"https://github.com/AIDajiangtang/Segment-Anything-CPP","commit_stats":null,"previous_names":["aidajiangtang/samtool-cpp","aidajiangtang/segment-anything-cpp"],"tags_count":0,"template":false,"template_full_name":null,"repository_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub/repositories/AIDajiangtang%2FSegment-Anything-CPP","tags_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub/repositories/AIDajiangtang%2FSegment-Anything-CPP/tags","releases_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub/repositories/AIDajiangtang%2FSegment-Anything-CPP/releases","manifests_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub/repositories/AIDajiangtang%2FSegment-Anything-CPP/manifests","owner_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub/owners/AIDajiangtang","download_url":"https://codeload.github.com/AIDajiangtang/Segment-Anything-CPP/tar.gz/refs/heads/main","host":{"name":"GitHub","url":"https://github.com","kind":"github","repositories_count":244310480,"owners_count":20432547,"icon_url":"https://github.com/github.png","version":null,"created_at":"2022-05-30T11:31:42.601Z","updated_at":"2022-07-04T15:15:14.044Z","host_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub","repositories_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub/repositories","repository_names_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub/repository_names","owners_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub/owners"}},"keywords":["deep-learning","segment-anything"],"created_at":"2024-07-31T03:00:51.622Z","updated_at":"2025-03-18T21:30:56.840Z","avatar_url":"https://github.com/AIDajiangtang.png","language":"C++","funding_links":[],"categories":["C++"],"sub_categories":[],"readme":"简体中文 | [English](ReadmeEN.md)  \n\n# segment anything（SAM）  \n[[`Paper`](https://ai.facebook.com/research/publications/segment-anything/)] [[`源码`](https://github.com/facebookresearch/segment-anything/)]  \n\n# Ours：segment anything（SAM） for CPP Inference  \n在SAMTool-CSharp仓库中，我们使用C#语言，ONNXRuntime for.Net框架对视觉大模型Segment Anaything完成了推理过程，并使用WPF与用户交互以及显示分割结果。  \n在本仓库SAMTool-CPP中，我们将使用C++语言，ONNXRuntime for CPP框架对视觉大模型Segment Anaything完成了推理过程，这样做出于两点考虑，第一，提升运行效率，第二，源码级款平台。  \n这样你就可以将模型部署到Windows，Linux，甚至Android等嵌入式设备中。  \nUI也不仅限于WPF了，你可以在QT，Html，Winform中任意选择，我们会提供WPF的版本。  \n\n微信公众号回复SAM获取第三方库：OpenCV和ONNXRuntime，然后将其解压到源码目录，OPenCV是我用Visual Studio2019编译的。ONNXRuntime不需要自己编译，下载官网编译好的即可  \n最终的项目结构如下：  \nproject_root/  \n  |- common.h  \n  |- CLIP.h  \n  |- CLIP.cpp  \n  |- SAM.h  \n  |- SAM.cpp \n  |- SimpleTokenizer.h \n  |- SimpleTokenizer.cpp \n  |- cppDemo.cpp  \n  |- opencv/  \n  |- onnxruntime-win-x64-1.15.1/  \n  |- CMakeLists.txt  \n\n\n# Text Promot：CLIP  \nSAM支持的promot有Point，Box，Mask以及Text文本，前三个相对比较容易实现和理解，Text Promot相对比较复杂。  \n对于Text Promot，需要引入另一个深度学习模型CLIP，简单来说，他能够将文本和图像映射到同一个向量空间中，这样文本和图像就能够进行比较了。   \n首先通过SAM分割出所有的目标，然后以每个目标为中心裁剪出一个图像，并将这个图像输入到CLIP中计算图像Embedding  \n然后通过CLIP计算输入文本的Embedding，最后计算文本Embedding和所有目标图像Embedding的余弦相似度，找到与文本最相似的目标图像。   \n","project_url":"https://awesome.ecosyste.ms/api/v1/projects/github.com%2FAIDajiangtang%2FSegment-Anything-CPP","html_url":"https://awesome.ecosyste.ms/projects/github.com%2FAIDajiangtang%2FSegment-Anything-CPP","lists_url":"https://awesome.ecosyste.ms/api/v1/projects/github.com%2FAIDajiangtang%2FSegment-Anything-CPP/lists"}