{"id":13627529,"url":"https://github.com/buptweixin/mxnet-deeplab","last_synced_at":"2025-04-17T00:31:50.721Z","repository":{"id":100714740,"uuid":"83308513","full_name":"buptweixin/mxnet-deeplab","owner":"buptweixin","description":"Deeplab for semantic segmentation implemented by MXNet ","archived":false,"fork":false,"pushed_at":"2017-03-05T14:17:08.000Z","size":25,"stargazers_count":22,"open_issues_count":2,"forks_count":4,"subscribers_count":3,"default_branch":"master","last_synced_at":"2024-11-08T18:44:50.787Z","etag":null,"topics":["deeplab","mxnet","semantic-segmentation"],"latest_commit_sha":null,"homepage":null,"language":"Python","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/buptweixin.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}},"created_at":"2017-02-27T12:35:25.000Z","updated_at":"2023-04-25T16:03:30.000Z","dependencies_parsed_at":null,"dependency_job_id":"c18a63eb-77aa-4145-8733-42f729d080a8","html_url":"https://github.com/buptweixin/mxnet-deeplab","commit_stats":null,"previous_names":[],"tags_count":0,"template":false,"template_full_name":null,"repository_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub/repositories/buptweixin%2Fmxnet-deeplab","tags_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub/repositories/buptweixin%2Fmxnet-deeplab/tags","releases_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub/repositories/buptweixin%2Fmxnet-deeplab/releases","manifests_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub/repositories/buptweixin%2Fmxnet-deeplab/manifests","owner_url":"https://repos.ecosyste.ms/api/v1/hosts/GitHub/owners/buptweixin","download_url":"https://codeload.github.com/buptweixin/mxnet-deeplab/tar.gz/refs/heads/master","host":{"name":"GitHub","url":"https://github.com","kind":"github","repositories_count":249292972,"owners_count":21245652,"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":["deeplab","mxnet","semantic-segmentation"],"created_at":"2024-08-01T22:00:35.216Z","updated_at":"2025-04-17T00:31:50.482Z","avatar_url":"https://github.com/buptweixin.png","language":"Python","funding_links":[],"categories":["\u003ca name=\"Vision\"\u003e\u003c/a\u003e2. Vision"],"sub_categories":["2.3 Image Segmentation"],"readme":"本项目将Caffe版本的DeepLab v2-LargeFOV网络移植到了MXNet框架.\n## 使用方法：\n### 依赖\n1. mxnet\n    version \u003e 0.7\n2. opencv及其python接口\n### 数据集准备\n从百度云下载预训练好的VGG16模型和参数文件到工程根目录:\n```shell\n# 下载VGG_FC_ILSVRC_16_layers-symbol.json\nwget --refer \"http://pan.baidu.com/s/1bgz4PC\" -O VGG_FC_ILSVRC_16_layers-symbol.json \"https://3grauymt1go3nhcttfa3ug.ourdvsss.com/d1.baidupcs.com/file/3990a272c33b0242f02420c9a130d640?bkt=p3-14003990a272c33b0242f02420c9a130d64079c68c30000000003b6a\u0026xcode=2bb72c1609f809d1916aec7057a1518716d653c658a55f46a7103330c9091c9b\u0026fid=1108131987-250528-1034373700851642\u0026time=1488202954\u0026sign=FDTAXGERLBH-DCb740ccc5511e5e8fedcff06b081203-3tYh9R%2F%2F7p9F3te2s1SPKRrY5%2B0%3D\u0026to=sf\u0026fm=Yan,B,U,nc\u0026sta_dx=15210\u0026sta_cs=350\u0026sta_ft=json\u0026sta_ct=7\u0026sta_mt=7\u0026fm2=Yangquan,B,U,nc\u0026newver=1\u0026newfm=1\u0026secfm=1\u0026flow_ver=3\u0026pkey=14003990a272c33b0242f02420c9a130d64079c68c30000000003b6a\u0026sl=72286287\u0026expires=8h\u0026rt=sh\u0026r=968934640\u0026mlogid=1342970328391683748\u0026vuk=3289204393\u0026vbdid=3039681765\u0026fin=VGG_FC_ILSVRC_16_layers-symbol.json\u0026fn=VGG_FC_ILSVRC_16_layers-symbol.json\u0026slt=pm\u0026uta=0\u0026rtype=1\u0026iv=0\u0026isw=0\u0026dp-logid=1342970328391683748\u0026dp-callid=0.1.1\u0026hps=1\u0026csl=400\u0026csign=NuOTkGygYoSPLqoaZF1HHMNTTIA%3D\u0026by=flowserver\u0026wshc_tag=0\u0026wsts_tag=58b42ccb\u0026wsid_tag=3d94f4ae\u0026wsiphost=ipdbm\"\n# 下载VGG_FC_ILSVR_16_layers-0074.params\nwget --refer \"http://pan.baidu.com/s/1bgz4PC\" -O VGG_FC_ILSVR_16_layers-0074.params \"https://qdcache00.baidupcs.com/file/4d805929e82225892ecbee68c33cc648?bkt=p3-0000d1495a68d33493685f8e663ddb61eb06\u0026xcode=2708ca2c3104a7a3eb1c30a0cadc72b6c34f7fb60676eabd1682cb8519c2059f\u0026fid=1108131987-250528-409289218452971\u0026time=1488203032\u0026sign=FDTAXGERLBH-DCb740ccc5511e5e8fedcff06b081203-gJocdHKzbBMjzJQE%2FenZMlZ%2BF%2BU%3D\u0026to=qd00\u0026fm=Nan,B,U,nc\u0026sta_dx=553431816\u0026sta_cs=41\u0026sta_ft=params\u0026sta_ct=7\u0026sta_mt=7\u0026fm2=Nanjing02,B,U,nc\u0026newver=1\u0026newfm=1\u0026secfm=1\u0026flow_ver=3\u0026pkey=0000d1495a68d33493685f8e663ddb61eb06\u0026sl=70123598\u0026expires=8h\u0026rt=sh\u0026r=344308576\u0026mlogid=1342991402114828163\u0026vuk=3289204393\u0026vbdid=3039681765\u0026fin=VGG_FC_ILSVRC_16_layers-0074.params\u0026fn=VGG_FC_ILSVRC_16_layers-0074.params\u0026slt=pm\u0026uta=0\u0026rtype=1\u0026iv=0\u0026isw=0\u0026dp-logid=1342991402114828163\u0026dp-callid=0.1.1\u0026hps=1\u0026csl=241\u0026csign=Zi9ouGXjhj3biOF08bDZFKVcEI8%3D\u0026by=flowserver\"\n```\n下载ISPRS数据集(ss)[https://pan.baidu.com/s/1kUQxs8J]\n```shell\nwget --refer \"https://pan.baidu.com/s/1kUQxs8J\" -O ISPRS.tar.gz \"https://nj02all01.baidupcs.com/file/e4c13a020304544513d29e545b18c688?bkt=p3-00001ad5dace852e8ae5fb4531387e8864df\u0026fid=3289204393-250528-459728222016\u0026time=1488204092\u0026sign=FDTAXGERLBH-DCb740ccc5511e5e8fedcff06b081203-epsiDCjUMkrcStwOefxj3RLQDbo%3D\u0026to=nj2hb\u0026fm=Yan,B,U,nc\u0026sta_dx=467793561\u0026sta_cs=\u0026sta_ft=gz\u0026sta_ct=0\u0026sta_mt=0\u0026fm2=Yangquan,B,U,nc\u0026newver=1\u0026newfm=1\u0026secfm=1\u0026flow_ver=3\u0026pkey=00001ad5dace852e8ae5fb4531387e8864df\u0026sl=76480590\u0026expires=8h\u0026rt=sh\u0026r=430772044\u0026mlogid=1343275688902104264\u0026vuk=3289204393\u0026vbdid=3039681765\u0026fin=ISPRS.tar.gz\u0026fn=ISPRS.tar.gz\u0026slt=pm\u0026uta=0\u0026rtype=1\u0026iv=0\u0026isw=0\u0026dp-logid=1343275688902104264\u0026dp-callid=0.1.1\u0026hps=1\u0026csl=80\u0026csign=27%2F%2BmaOjbfOdfUYOWleCxsS4EqM%3D\u0026by=flowserver\"\n# 解压数据集到工程根目录\ntar -xzvf ISPRS.tar.gz\n```\n在`run_deeplab.sh`中设置数据集根路径，比如\"root_dir=./ISPRS/\"\n\n### 运行\n```shell\n./run_deeplab.sh\n```\n部分日志\n```shell\nINFO:root:Start training with gpu(0)\nINFO:root:Epoch[0] Batch [10]   Speed: 5.68 samples/sec Train-accuracy=0.781917\nINFO:root:Epoch[0] Batch [20]   Speed: 5.17 samples/sec Train-accuracy=0.789799\nINFO:root:Epoch[0] Batch [30]   Speed: 5.17 samples/sec Train-accuracy=0.772426\nINFO:root:Epoch[0] Batch [40]   Speed: 5.17 samples/sec Train-accuracy=0.806343\nINFO:root:Epoch[0] Batch [50]   Speed: 5.18 samples/sec Train-accuracy=0.847787\nINFO:root:Epoch[0] Batch [60]   Speed: 5.18 samples/sec Train-accuracy=0.827195\nINFO:root:Epoch[0] Batch [70]   Speed: 5.19 samples/sec Train-accuracy=0.809657\n...\n...\nINFO:root:Epoch[49] Batch [740] Speed: 5.13 samples/sec Train-accuracy=0.965917\nINFO:root:Epoch[49] Batch [750] Speed: 5.13 samples/sec Train-accuracy=0.958509\nINFO:root:Epoch[49] Batch [760] Speed: 5.12 samples/sec Train-accuracy=0.965964\nINFO:root:Epoch[49] Batch [770] Speed: 5.13 samples/sec Train-accuracy=0.962391\nINFO:root:Saved checkpoint to \"DeepLab-V2-0050.params\"\nINFO:root:                            ---\u003eEpoch[49] Train-accuracy=0.968494\nINFO:root: in eval process...\nINFO:root:batch[179] Validation-accuracy=0.837493\n```\n\n### 效果\n- 原图\n\n    ![原图](http://ww1.sinaimg.cn/large/6425ef91ly1fd5e8a7ulaj20eo0e8tle)\n- 效果图\n\n    ![网络输出](http://ww1.sinaimg.cn/large/6425ef91ly1fd5e6xfz9oj20e70e4aag)\n","project_url":"https://awesome.ecosyste.ms/api/v1/projects/github.com%2Fbuptweixin%2Fmxnet-deeplab","html_url":"https://awesome.ecosyste.ms/projects/github.com%2Fbuptweixin%2Fmxnet-deeplab","lists_url":"https://awesome.ecosyste.ms/api/v1/projects/github.com%2Fbuptweixin%2Fmxnet-deeplab/lists"}