{"version":2,"relationshipURLBase":"https://codingai-lec03.pages.dev/course/study/pixels-instances.html","relationshipResolution":"Resolve relative relationship URLs against relationshipURLBase; complete records and absolute URLs are in blocks.","title":"像素、边界与对象身份","url":"https://codingai-lec03.pages.dev/course/study/pixels-instances.html","markdown":"https://codingai-lec03.pages.dev/course/study/pixels-instances.md","pages":[{"id":"reading-opening","title":"像素、边界与对象身份","unit":"像素、边界与对象身份","unitId":"study/pixels-instances","chapter":"course","number":1},{"id":"reading-steps","title":"阅读与动手顺序","unit":"像素、边界与对象身份","unitId":"study/pixels-instances","chapter":"course","number":2},{"id":"reading-output","title":"带走的结果","unit":"像素、边界与对象身份","unitId":"study/pixels-instances","chapter":"course","number":3},{"id":"reading-practice","title":"优先实践","unit":"像素、边界与对象身份","unitId":"study/pixels-instances","chapter":"course","number":4},{"id":"reading-readings","title":"配套讲解","unit":"像素、边界与对象身份","unitId":"study/pixels-instances","chapter":"course","number":5},{"id":"reading-extensions","title":"遇到相应问题时再展开（3 项）","unit":"像素、边界与对象身份","unitId":"study/pixels-instances","chapter":"course","number":6},{"id":"reading-classroom","title":"接回课堂中的问题","unit":"像素、边界与对象身份","unitId":"study/pixels-instances","chapter":"course","number":7}],"aliases":[{"oldId":"b-bd1328993a28836bfffb13f3","targetId":"b-bd1328993a28836bfffb13f3","page":"reading-opening","operation":"move","beforeSHA256":"b31d47353725d79028b93a0678b34bbadb488696e9b6fec06c2641086f611773","afterSHA256":"b31d47353725d79028b93a0678b34bbadb488696e9b6fec06c2641086f611773","reason":"Same source body in authored semantic page"},{"oldId":"b-339499dc67bdd1417210e097","targetId":"b-339499dc67bdd1417210e097","page":"reading-opening","operation":"move","beforeSHA256":"422d4b4e59346648c07d21784f2e84395facc6310a1241a3ccab6281f348aaca","afterSHA256":"422d4b4e59346648c07d21784f2e84395facc6310a1241a3ccab6281f348aaca","reason":"Same source body in authored semantic page"},{"oldId":"b-26ab46edebaa80ef40af1bb4","targetId":"b-26ab46edebaa80ef40af1bb4","page":"reading-steps","operation":"move","beforeSHA256":"fe70d0188def3730a8b7b86505cbbe3b1936f75fbed253c2d4e637a968cfe45b","afterSHA256":"fe70d0188def3730a8b7b86505cbbe3b1936f75fbed253c2d4e637a968cfe45b","reason":"Same source body in authored semantic page"},{"oldId":"b-a92a807c335fcc4eb2b73430","targetId":"b-a92a807c335fcc4eb2b73430","page":"reading-steps","operation":"move","beforeSHA256":"3d28e664b869acdd8c725d9fa4fecb07437c604f8ad80cf9bfd27ab3624d34d4","afterSHA256":"3d28e664b869acdd8c725d9fa4fecb07437c604f8ad80cf9bfd27ab3624d34d4","reason":"Same source body in authored semantic page"},{"oldId":"b-f6cba0e8dac0f36ac6034ef9","targetId":"b-f6cba0e8dac0f36ac6034ef9","page":"reading-steps","operation":"move","beforeSHA256":"26c7c5a7e5bb7864d9a462559607bb5d5781e1b7e368b042a5475af9737a7c8f","afterSHA256":"26c7c5a7e5bb7864d9a462559607bb5d5781e1b7e368b042a5475af9737a7c8f","reason":"Same source body in authored semantic page"},{"oldId":"b-7aa261e4827fb48da8276497","targetId":"b-7aa261e4827fb48da8276497","page":"reading-steps","operation":"move","beforeSHA256":"2542f940fdf58ed9c3f7e46645779f05109a616364b8663d6b57111b1e62bb2b","afterSHA256":"2542f940fdf58ed9c3f7e46645779f05109a616364b8663d6b57111b1e62bb2b","reason":"Same source body in authored semantic page"},{"oldId":"b-dab32a049bed7bdfb96fa6a3","targetId":"b-dab32a049bed7bdfb96fa6a3","page":"reading-steps","operation":"move","beforeSHA256":"7c18d1c61b0ffde5298419806f9debd0dcd51d31f6b0ef093ed1fab2e3735834","afterSHA256":"7c18d1c61b0ffde5298419806f9debd0dcd51d31f6b0ef093ed1fab2e3735834","reason":"Same source body in authored semantic page"},{"oldId":"b-a2363ddcbae5d61c53d6ed7a","targetId":"b-a2363ddcbae5d61c53d6ed7a","page":"reading-steps","operation":"move","beforeSHA256":"2fd54d1ec0d17ed5fc6705852e84e4ff9fa22092ffb71bfb4195955340452289","afterSHA256":"2fd54d1ec0d17ed5fc6705852e84e4ff9fa22092ffb71bfb4195955340452289","reason":"Same source body in authored semantic page"},{"oldId":"b-ad8c95694a3015651ebd8a0f","targetId":"b-ad8c95694a3015651ebd8a0f","page":"reading-steps","operation":"move","beforeSHA256":"e074dd870cb835d2a74c2e80286af1f331b6e01f75e6ccc122c92aba9ae32854","afterSHA256":"e074dd870cb835d2a74c2e80286af1f331b6e01f75e6ccc122c92aba9ae32854","reason":"Same source body in authored semantic page"},{"oldId":"b-35537419550ce5b3faebe731","targetId":"b-35537419550ce5b3faebe731","page":"reading-steps","operation":"move","beforeSHA256":"42c1daaa7fdcd8724cd3b057d4e5bd79e6ede88cc843aeed1052325def1b85e2","afterSHA256":"42c1daaa7fdcd8724cd3b057d4e5bd79e6ede88cc843aeed1052325def1b85e2","reason":"Same source body in authored semantic page"},{"oldId":"b-14901007f0b9afa5ee031daf","targetId":"b-14901007f0b9afa5ee031daf","page":"reading-steps","operation":"move","beforeSHA256":"dc47c411d7a5907c6970c1f0edbebaf15496a7f539ee3db3820beeac45d7874a","afterSHA256":"dc47c411d7a5907c6970c1f0edbebaf15496a7f539ee3db3820beeac45d7874a","reason":"Same source body in authored semantic page"},{"oldId":"b-b819b99183515f3d9013b3b9","targetId":"b-b819b99183515f3d9013b3b9","page":"reading-steps","operation":"move","beforeSHA256":"93474f2f2b9c40c18d2bfd508afe9a8a7952694e84a0ffe8f52315006631bc8f","afterSHA256":"93474f2f2b9c40c18d2bfd508afe9a8a7952694e84a0ffe8f52315006631bc8f","reason":"Same source body in authored semantic page"},{"oldId":"b-6fa06ceb500b5b01c5b999ad","targetId":"b-6fa06ceb500b5b01c5b999ad","page":"reading-steps","operation":"move","beforeSHA256":"5eb8e5795a8b7f92056cc4efe2cfee45d7e5033d089b2a7722dd92c79abd948b","afterSHA256":"5eb8e5795a8b7f92056cc4efe2cfee45d7e5033d089b2a7722dd92c79abd948b","reason":"Same source body in authored semantic page"},{"oldId":"b-a90e49fc60c84e196a618b65","targetId":"b-a90e49fc60c84e196a618b65","page":"reading-output","operation":"move","beforeSHA256":"b36b36fed7907fe618dc125fe4123b9f998461c6dab4ab8dfbfdfc4c94c8cc25","afterSHA256":"b36b36fed7907fe618dc125fe4123b9f998461c6dab4ab8dfbfdfc4c94c8cc25","reason":"Same source body in authored semantic page"},{"oldId":"b-23b4a16aa04a17c0d826b97e","targetId":"b-23b4a16aa04a17c0d826b97e","page":"reading-output","operation":"move","beforeSHA256":"6df7f518eaa6d443a75391daef818bd7331b266e8cc60c912cac745e6f06e95b","afterSHA256":"6df7f518eaa6d443a75391daef818bd7331b266e8cc60c912cac745e6f06e95b","reason":"Same source body in authored semantic page"},{"oldId":"b-e14129aa4f8e423abbda0afe","targetId":"b-e14129aa4f8e423abbda0afe","page":"reading-practice","operation":"move","beforeSHA256":"0b242f211890b34b91f0dce9c1b60fa872957d318ee44d54c0ea6566dd831ab9","afterSHA256":"0b242f211890b34b91f0dce9c1b60fa872957d318ee44d54c0ea6566dd831ab9","reason":"Same source body in authored semantic page"},{"oldId":"b-47e55ed019bedf0e240ae160","targetId":"b-47e55ed019bedf0e240ae160","page":"reading-practice","operation":"move","beforeSHA256":"19608cacbee0cd46460095bdebd3ee178d3479857b413a9a631e4b79f688aaef","afterSHA256":"19608cacbee0cd46460095bdebd3ee178d3479857b413a9a631e4b79f688aaef","reason":"Same source body in authored semantic page"},{"oldId":"b-294c54ac97501a67bb5353ac","targetId":"b-294c54ac97501a67bb5353ac","page":"reading-practice","operation":"move","beforeSHA256":"ad954fa60070acddb3217b443b31bbcd49b21f18d1e8ff02e33fd0eb3319879d","afterSHA256":"ad954fa60070acddb3217b443b31bbcd49b21f18d1e8ff02e33fd0eb3319879d","reason":"Same source body in authored semantic page"},{"oldId":"b-31608ee2012235246a30dc36","targetId":"b-31608ee2012235246a30dc36","page":"reading-practice","operation":"move","beforeSHA256":"ad954fa60070acddb3217b443b31bbcd49b21f18d1e8ff02e33fd0eb3319879d","afterSHA256":"ad954fa60070acddb3217b443b31bbcd49b21f18d1e8ff02e33fd0eb3319879d","reason":"Same source body in authored semantic page"},{"oldId":"b-fbd0e43cca03697df4366716","targetId":"b-fbd0e43cca03697df4366716","page":"reading-practice","operation":"move","beforeSHA256":"6f0b115c3c32032428a4c0d5ec4791e16d6b63aa3d643457bed758bf6af2b765","afterSHA256":"6f0b115c3c32032428a4c0d5ec4791e16d6b63aa3d643457bed758bf6af2b765","reason":"Same source body in authored semantic page"},{"oldId":"b-54cff3ad2eafb279b968775b","targetId":"b-54cff3ad2eafb279b968775b","page":"reading-readings","operation":"move","beforeSHA256":"1c792c1a05326d4a07eba3b245e97c5be88fcadf7f81a2700ca1963b4c7f0299","afterSHA256":"1c792c1a05326d4a07eba3b245e97c5be88fcadf7f81a2700ca1963b4c7f0299","reason":"Same source body in authored semantic page"},{"oldId":"b-506825ef2fdd72bf477f7e82","targetId":"b-506825ef2fdd72bf477f7e82","page":"reading-readings","operation":"move","beforeSHA256":"291a8d14abd034c805ac2152e5c91f4d1fa63c32680ee4d789cfcf3aa83b3907","afterSHA256":"291a8d14abd034c805ac2152e5c91f4d1fa63c32680ee4d789cfcf3aa83b3907","reason":"Same source body in authored semantic page"},{"oldId":"b-5a75e3d4ca42f57cf72af22e","targetId":"b-5a75e3d4ca42f57cf72af22e","page":"reading-readings","operation":"move","beforeSHA256":"fb96de8722fe7f9734018668e519678db20211cf892e7b05cc571067ea9c8b10","afterSHA256":"fb96de8722fe7f9734018668e519678db20211cf892e7b05cc571067ea9c8b10","reason":"Same source body in authored semantic page"},{"oldId":"b-d74db135aa33e8baad5ff504","targetId":"b-d74db135aa33e8baad5ff504","page":"reading-readings","operation":"move","beforeSHA256":"e0584a8ed4ddfb5b9c9abb132012919eb506480c39698c3ddd67fccd73d0a12e","afterSHA256":"e0584a8ed4ddfb5b9c9abb132012919eb506480c39698c3ddd67fccd73d0a12e","reason":"Same source body in authored semantic page"},{"oldId":"b-4eef245e5ad6cdb7b9492a66","targetId":"b-4eef245e5ad6cdb7b9492a66","page":"reading-readings","operation":"move","beforeSHA256":"13fd7efed3962e7ae1181a1c43b5745a38874692751a3ee2e455036c500ef0ff","afterSHA256":"13fd7efed3962e7ae1181a1c43b5745a38874692751a3ee2e455036c500ef0ff","reason":"Same source body in authored semantic page"},{"oldId":"b-d6eb1414ec3969b3d53f344e","targetId":"b-d6eb1414ec3969b3d53f344e","page":"reading-readings","operation":"move","beforeSHA256":"70eec56f79b131d4695a1c420bc695a02ef3863727481994de8f7ed4bd673dcc","afterSHA256":"70eec56f79b131d4695a1c420bc695a02ef3863727481994de8f7ed4bd673dcc","reason":"Same source body in authored semantic page"},{"oldId":"b-b1dc84b1acfe27ad6bef2131","targetId":"b-b1dc84b1acfe27ad6bef2131","page":"reading-extensions","operation":"move","beforeSHA256":"e79866b4a5a13c04e04af53c771cb58efafc6484194549655394bc08f209a0cc","afterSHA256":"e79866b4a5a13c04e04af53c771cb58efafc6484194549655394bc08f209a0cc","reason":"Same source body in authored semantic page"},{"oldId":"b-ea92cc7e755ea10d513af8e4","targetId":"b-ea92cc7e755ea10d513af8e4","page":"reading-extensions","operation":"move","beforeSHA256":"23e6a723a12be191ae6889710e28c2341e839e8a0af1450d2c724a6d308b233c","afterSHA256":"23e6a723a12be191ae6889710e28c2341e839e8a0af1450d2c724a6d308b233c","reason":"Same source body in authored semantic page"},{"oldId":"b-f17b5385818f75a2df4f3db4","targetId":"b-f17b5385818f75a2df4f3db4","page":"reading-extensions","operation":"move","beforeSHA256":"93fe7c165c595944c7af5b3abd9e50016526d59fcb6d412fbfb865bd208f77c4","afterSHA256":"93fe7c165c595944c7af5b3abd9e50016526d59fcb6d412fbfb865bd208f77c4","reason":"Same source body in authored semantic page"},{"oldId":"b-b378f89fc7d41f333daf2065","targetId":"b-b378f89fc7d41f333daf2065","page":"reading-extensions","operation":"move","beforeSHA256":"86302464db105c037b947be3a1cd9e3448d76128b7a7dfdd261967230ef4cd2b","afterSHA256":"86302464db105c037b947be3a1cd9e3448d76128b7a7dfdd261967230ef4cd2b","reason":"Same source body in authored semantic page"},{"oldId":"b-4ad6607619c35a6f8bca6059","targetId":"b-4ad6607619c35a6f8bca6059","page":"reading-extensions","operation":"move","beforeSHA256":"7d1e7531a3ac0397d25516d56db48a938b043ef98467022f8eacc41563e73771","afterSHA256":"7d1e7531a3ac0397d25516d56db48a938b043ef98467022f8eacc41563e73771","reason":"Same source body in authored semantic page"},{"oldId":"b-7a420a4e5aabcac8fa1961bf","targetId":"b-7a420a4e5aabcac8fa1961bf","page":"reading-extensions","operation":"move","beforeSHA256":"fd816034dc790e92c947b11e0bc8f6df6adb8394c5bd8d086b24a71fc6a60530","afterSHA256":"fd816034dc790e92c947b11e0bc8f6df6adb8394c5bd8d086b24a71fc6a60530","reason":"Same source body in authored semantic page"},{"oldId":"b-a67a92515ed378b5d611b77c","targetId":"b-a67a92515ed378b5d611b77c","page":"reading-extensions","operation":"move","beforeSHA256":"1d7535d0c02cc7bae4e9740472c8fa2e46f9a3b9067705762875bccfe9114541","afterSHA256":"1d7535d0c02cc7bae4e9740472c8fa2e46f9a3b9067705762875bccfe9114541","reason":"Same source body in authored semantic page"},{"oldId":"b-ee4ba5df9fa064e1db1dbba6","targetId":"b-ee4ba5df9fa064e1db1dbba6","page":"reading-classroom","operation":"move","beforeSHA256":"5214b785a21ffd090bfef76fc13e4d2fd2444ca45eccd08a1f2b475365a5634a","afterSHA256":"5214b785a21ffd090bfef76fc13e4d2fd2444ca45eccd08a1f2b475365a5634a","reason":"Same source body in authored semantic page"}],"blocks":[{"id":"b-bd1328993a28836bfffb13f3","type":"h1","scope":"document","url":"https://codingai-lec03.pages.dev/course/study/pixels-instances.html#b-bd1328993a28836bfffb13f3","content":"# 像素、边界与对象身份","sha256":"b31d47353725d79028b93a0678b34bbadb488696e9b6fec06c2641086f611773","chapter":{"id":"course","title":"像素、边界与对象身份"},"unit":{"id":"study/pixels-instances","title":"像素、边界与对象身份"},"page":{"id":"reading-opening","title":"像素、边界与对象身份","number":1},"sectionTitle":"像素、边界与对象身份","currentQuestion":"像素、边界与对象身份","sourceSlides":[],"knowledgeBlock":null,"formulas":[],"previousBlock":null,"nextBlock":"b-339499dc67bdd1417210e097","prerequisites":[{"id":"b-339499dc67bdd1417210e097","url":"#b-339499dc67bdd1417210e097","role":"problem-context"}],"relatedBlocks":[]},{"id":"b-339499dc67bdd1417210e097","type":"p","scope":"document","url":"https://codingai-lec03.pages.dev/course/study/pixels-instances.html#b-339499dc67bdd1417210e097","content":"输出格子变多之后，模型怎样知道每个位置属于什么？","sha256":"422d4b4e59346648c07d21784f2e84395facc6310a1241a3ccab6281f348aaca","chapter":{"id":"course","title":"像素、边界与对象身份"},"unit":{"id":"study/pixels-instances","title":"像素、边界与对象身份"},"page":{"id":"reading-opening","title":"像素、边界与对象身份","number":1},"sectionTitle":"像素、边界与对象身份","currentQuestion":"像素、边界与对象身份","sourceSlides":[],"knowledgeBlock":null,"formulas":[],"previousBlock":"b-bd1328993a28836bfffb13f3","nextBlock":null,"prerequisites":[{"id":"b-bd1328993a28836bfffb13f3","url":"#b-bd1328993a28836bfffb13f3","role":"section-title"}],"relatedBlocks":[]},{"id":"b-26ab46edebaa80ef40af1bb4","type":"h2","scope":"steps","url":"https://codingai-lec03.pages.dev/course/study/pixels-instances.html#b-26ab46edebaa80ef40af1bb4","content":"## 阅读与动手顺序","sha256":"fe70d0188def3730a8b7b86505cbbe3b1936f75fbed253c2d4e637a968cfe45b","chapter":{"id":"course","title":"像素、边界与对象身份"},"unit":{"id":"study/pixels-instances","title":"像素、边界与对象身份"},"page":{"id":"reading-steps","title":"阅读与动手顺序","number":2},"sectionTitle":"阅读与动手顺序","currentQuestion":"阅读与动手顺序","sourceSlides":[],"knowledgeBlock":null,"formulas":[],"previousBlock":null,"nextBlock":"b-a92a807c335fcc4eb2b73430","prerequisites":[{"id":"b-dab32a049bed7bdfb96fa6a3","url":"#b-dab32a049bed7bdfb96fa6a3","role":"problem-context"}],"relatedBlocks":[]},{"id":"b-a92a807c335fcc4eb2b73430","type":"ol","scope":"steps","url":"https://codingai-lec03.pages.dev/course/study/pixels-instances.html#b-a92a807c335fcc4eb2b73430","content":"- ### 区分采样练习与像素学习\n\nLAB03 解释地址、插值与信息碰撞；LAB11 才比较从 RGB 预测类别 mask 的学习模型。\n- ### 读同一任务的模型对照\n\n比较简单 RGB 读出与 FCN 的预测、错误图及逐类 IoU，并检查模糊后两者如何变化。\n- ### 按需要增加对象条件\n\n若问题变成“选哪一个对象”，再到 LAB16 看正负点与框提示的规则机制。它不使用 SAM 权重。","sha256":"3d28e664b869acdd8c725d9fa4fecb07437c604f8ad80cf9bfd27ab3624d34d4","chapter":{"id":"course","title":"像素、边界与对象身份"},"unit":{"id":"study/pixels-instances","title":"像素、边界与对象身份"},"page":{"id":"reading-steps","title":"阅读与动手顺序","number":2},"sectionTitle":"阅读与动手顺序","currentQuestion":"阅读与动手顺序","sourceSlides":[],"knowledgeBlock":null,"formulas":[],"previousBlock":"b-26ab46edebaa80ef40af1bb4","nextBlock":"b-f6cba0e8dac0f36ac6034ef9","prerequisites":[{"id":"b-26ab46edebaa80ef40af1bb4","url":"#b-26ab46edebaa80ef40af1bb4","role":"section-title"},{"id":"b-dab32a049bed7bdfb96fa6a3","url":"#b-dab32a049bed7bdfb96fa6a3","role":"problem-context"}],"relatedBlocks":[]},{"id":"b-f6cba0e8dac0f36ac6034ef9","type":"li","scope":"steps","url":"https://codingai-lec03.pages.dev/course/study/pixels-instances.html#b-f6cba0e8dac0f36ac6034ef9","content":"- ### 区分采样练习与像素学习\n\nLAB03 解释地址、插值与信息碰撞；LAB11 才比较从 RGB 预测类别 mask 的学习模型。","sha256":"26c7c5a7e5bb7864d9a462559607bb5d5781e1b7e368b042a5475af9737a7c8f","chapter":{"id":"course","title":"像素、边界与对象身份"},"unit":{"id":"study/pixels-instances","title":"像素、边界与对象身份"},"page":{"id":"reading-steps","title":"阅读与动手顺序","number":2},"sectionTitle":"阅读与动手顺序","currentQuestion":"阅读与动手顺序","sourceSlides":[],"knowledgeBlock":null,"formulas":[],"previousBlock":"b-a92a807c335fcc4eb2b73430","nextBlock":"b-7aa261e4827fb48da8276497","prerequisites":[{"id":"b-26ab46edebaa80ef40af1bb4","url":"#b-26ab46edebaa80ef40af1bb4","role":"section-title"},{"id":"b-dab32a049bed7bdfb96fa6a3","url":"#b-dab32a049bed7bdfb96fa6a3","role":"problem-context"}],"relatedBlocks":[]},{"id":"b-7aa261e4827fb48da8276497","type":"h3","scope":"steps","url":"https://codingai-lec03.pages.dev/course/study/pixels-instances.html#b-7aa261e4827fb48da8276497","content":"### 区分采样练习与像素学习","sha256":"2542f940fdf58ed9c3f7e46645779f05109a616364b8663d6b57111b1e62bb2b","chapter":{"id":"course","title":"像素、边界与对象身份"},"unit":{"id":"study/pixels-instances","title":"像素、边界与对象身份"},"page":{"id":"reading-steps","title":"阅读与动手顺序","number":2},"sectionTitle":"阅读与动手顺序","currentQuestion":"阅读与动手顺序","sourceSlides":[],"knowledgeBlock":null,"formulas":[],"previousBlock":"b-f6cba0e8dac0f36ac6034ef9","nextBlock":"b-dab32a049bed7bdfb96fa6a3","prerequisites":[{"id":"b-26ab46edebaa80ef40af1bb4","url":"#b-26ab46edebaa80ef40af1bb4","role":"section-title"},{"id":"b-dab32a049bed7bdfb96fa6a3","url":"#b-dab32a049bed7bdfb96fa6a3","role":"problem-context"}],"relatedBlocks":[]},{"id":"b-dab32a049bed7bdfb96fa6a3","type":"p","scope":"steps","url":"https://codingai-lec03.pages.dev/course/study/pixels-instances.html#b-dab32a049bed7bdfb96fa6a3","content":"LAB03 解释地址、插值与信息碰撞；LAB11 才比较从 RGB 预测类别 mask 的学习模型。","sha256":"7c18d1c61b0ffde5298419806f9debd0dcd51d31f6b0ef093ed1fab2e3735834","chapter":{"id":"course","title":"像素、边界与对象身份"},"unit":{"id":"study/pixels-instances","title":"像素、边界与对象身份"},"page":{"id":"reading-steps","title":"阅读与动手顺序","number":2},"sectionTitle":"阅读与动手顺序","currentQuestion":"阅读与动手顺序","sourceSlides":[],"knowledgeBlock":null,"formulas":[],"previousBlock":"b-7aa261e4827fb48da8276497","nextBlock":"b-a2363ddcbae5d61c53d6ed7a","prerequisites":[{"id":"b-7aa261e4827fb48da8276497","url":"#b-7aa261e4827fb48da8276497","role":"section-title"}],"relatedBlocks":[]},{"id":"b-a2363ddcbae5d61c53d6ed7a","type":"li","scope":"steps","url":"https://codingai-lec03.pages.dev/course/study/pixels-instances.html#b-a2363ddcbae5d61c53d6ed7a","content":"- ### 读同一任务的模型对照\n\n比较简单 RGB 读出与 FCN 的预测、错误图及逐类 IoU，并检查模糊后两者如何变化。","sha256":"2fd54d1ec0d17ed5fc6705852e84e4ff9fa22092ffb71bfb4195955340452289","chapter":{"id":"course","title":"像素、边界与对象身份"},"unit":{"id":"study/pixels-instances","title":"像素、边界与对象身份"},"page":{"id":"reading-steps","title":"阅读与动手顺序","number":2},"sectionTitle":"阅读与动手顺序","currentQuestion":"阅读与动手顺序","sourceSlides":[],"knowledgeBlock":null,"formulas":[],"previousBlock":"b-dab32a049bed7bdfb96fa6a3","nextBlock":"b-ad8c95694a3015651ebd8a0f","prerequisites":[{"id":"b-7aa261e4827fb48da8276497","url":"#b-7aa261e4827fb48da8276497","role":"section-title"},{"id":"b-dab32a049bed7bdfb96fa6a3","url":"#b-dab32a049bed7bdfb96fa6a3","role":"problem-context"}],"relatedBlocks":[]},{"id":"b-ad8c95694a3015651ebd8a0f","type":"h3","scope":"steps","url":"https://codingai-lec03.pages.dev/course/study/pixels-instances.html#b-ad8c95694a3015651ebd8a0f","content":"### 读同一任务的模型对照","sha256":"e074dd870cb835d2a74c2e80286af1f331b6e01f75e6ccc122c92aba9ae32854","chapter":{"id":"course","title":"像素、边界与对象身份"},"unit":{"id":"study/pixels-instances","title":"像素、边界与对象身份"},"page":{"id":"reading-steps","title":"阅读与动手顺序","number":2},"sectionTitle":"阅读与动手顺序","currentQuestion":"阅读与动手顺序","sourceSlides":[],"knowledgeBlock":null,"formulas":[],"previousBlock":"b-a2363ddcbae5d61c53d6ed7a","nextBlock":"b-35537419550ce5b3faebe731","prerequisites":[{"id":"b-7aa261e4827fb48da8276497","url":"#b-7aa261e4827fb48da8276497","role":"section-title"},{"id":"b-dab32a049bed7bdfb96fa6a3","url":"#b-dab32a049bed7bdfb96fa6a3","role":"problem-context"}],"relatedBlocks":[]},{"id":"b-35537419550ce5b3faebe731","type":"p","scope":"steps","url":"https://codingai-lec03.pages.dev/course/study/pixels-instances.html#b-35537419550ce5b3faebe731","content":"比较简单 RGB 读出与 FCN 的预测、错误图及逐类 IoU，并检查模糊后两者如何变化。","sha256":"42c1daaa7fdcd8724cd3b057d4e5bd79e6ede88cc843aeed1052325def1b85e2","chapter":{"id":"course","title":"像素、边界与对象身份"},"unit":{"id":"study/pixels-instances","title":"像素、边界与对象身份"},"page":{"id":"reading-steps","title":"阅读与动手顺序","number":2},"sectionTitle":"阅读与动手顺序","currentQuestion":"阅读与动手顺序","sourceSlides":[],"knowledgeBlock":null,"formulas":[],"previousBlock":"b-ad8c95694a3015651ebd8a0f","nextBlock":"b-14901007f0b9afa5ee031daf","prerequisites":[{"id":"b-ad8c95694a3015651ebd8a0f","url":"#b-ad8c95694a3015651ebd8a0f","role":"section-title"},{"id":"b-dab32a049bed7bdfb96fa6a3","url":"#b-dab32a049bed7bdfb96fa6a3","role":"problem-context"}],"relatedBlocks":[]},{"id":"b-14901007f0b9afa5ee031daf","type":"li","scope":"steps","url":"https://codingai-lec03.pages.dev/course/study/pixels-instances.html#b-14901007f0b9afa5ee031daf","content":"- ### 按需要增加对象条件\n\n若问题变成“选哪一个对象”，再到 LAB16 看正负点与框提示的规则机制。它不使用 SAM 权重。","sha256":"dc47c411d7a5907c6970c1f0edbebaf15496a7f539ee3db3820beeac45d7874a","chapter":{"id":"course","title":"像素、边界与对象身份"},"unit":{"id":"study/pixels-instances","title":"像素、边界与对象身份"},"page":{"id":"reading-steps","title":"阅读与动手顺序","number":2},"sectionTitle":"阅读与动手顺序","currentQuestion":"阅读与动手顺序","sourceSlides":[],"knowledgeBlock":null,"formulas":[],"previousBlock":"b-35537419550ce5b3faebe731","nextBlock":"b-b819b99183515f3d9013b3b9","prerequisites":[{"id":"b-ad8c95694a3015651ebd8a0f","url":"#b-ad8c95694a3015651ebd8a0f","role":"section-title"},{"id":"b-dab32a049bed7bdfb96fa6a3","url":"#b-dab32a049bed7bdfb96fa6a3","role":"problem-context"}],"relatedBlocks":[]},{"id":"b-b819b99183515f3d9013b3b9","type":"h3","scope":"steps","url":"https://codingai-lec03.pages.dev/course/study/pixels-instances.html#b-b819b99183515f3d9013b3b9","content":"### 按需要增加对象条件","sha256":"93474f2f2b9c40c18d2bfd508afe9a8a7952694e84a0ffe8f52315006631bc8f","chapter":{"id":"course","title":"像素、边界与对象身份"},"unit":{"id":"study/pixels-instances","title":"像素、边界与对象身份"},"page":{"id":"reading-steps","title":"阅读与动手顺序","number":2},"sectionTitle":"阅读与动手顺序","currentQuestion":"阅读与动手顺序","sourceSlides":[],"knowledgeBlock":null,"formulas":[],"previousBlock":"b-14901007f0b9afa5ee031daf","nextBlock":"b-6fa06ceb500b5b01c5b999ad","prerequisites":[{"id":"b-ad8c95694a3015651ebd8a0f","url":"#b-ad8c95694a3015651ebd8a0f","role":"section-title"},{"id":"b-dab32a049bed7bdfb96fa6a3","url":"#b-dab32a049bed7bdfb96fa6a3","role":"problem-context"}],"relatedBlocks":[]},{"id":"b-6fa06ceb500b5b01c5b999ad","type":"p","scope":"steps","url":"https://codingai-lec03.pages.dev/course/study/pixels-instances.html#b-6fa06ceb500b5b01c5b999ad","content":"若问题变成“选哪一个对象”，再到 LAB16 看正负点与框提示的规则机制。它不使用 SAM 权重。","sha256":"5eb8e5795a8b7f92056cc4efe2cfee45d7e5033d089b2a7722dd92c79abd948b","chapter":{"id":"course","title":"像素、边界与对象身份"},"unit":{"id":"study/pixels-instances","title":"像素、边界与对象身份"},"page":{"id":"reading-steps","title":"阅读与动手顺序","number":2},"sectionTitle":"阅读与动手顺序","currentQuestion":"阅读与动手顺序","sourceSlides":[],"knowledgeBlock":null,"formulas":[],"previousBlock":"b-b819b99183515f3d9013b3b9","nextBlock":null,"prerequisites":[{"id":"b-b819b99183515f3d9013b3b9","url":"#b-b819b99183515f3d9013b3b9","role":"section-title"},{"id":"b-dab32a049bed7bdfb96fa6a3","url":"#b-dab32a049bed7bdfb96fa6a3","role":"problem-context"}],"relatedBlocks":[]},{"id":"b-a90e49fc60c84e196a618b65","type":"h2","scope":"output","url":"https://codingai-lec03.pages.dev/course/study/pixels-instances.html#b-a90e49fc60c84e196a618b65","content":"## 带走的结果","sha256":"b36b36fed7907fe618dc125fe4123b9f998461c6dab4ab8dfbfdfc4c94c8cc25","chapter":{"id":"course","title":"像素、边界与对象身份"},"unit":{"id":"study/pixels-instances","title":"像素、边界与对象身份"},"page":{"id":"reading-output","title":"带走的结果","number":3},"sectionTitle":"带走的结果","currentQuestion":"带走的结果","sourceSlides":[],"knowledgeBlock":null,"formulas":[],"previousBlock":null,"nextBlock":"b-23b4a16aa04a17c0d826b97e","prerequisites":[{"id":"b-23b4a16aa04a17c0d826b97e","url":"#b-23b4a16aa04a17c0d826b97e","role":"problem-context"}],"relatedBlocks":[]},{"id":"b-23b4a16aa04a17c0d826b97e","type":"p","scope":"output","url":"https://codingai-lec03.pages.dev/course/study/pixels-instances.html#b-23b4a16aa04a17c0d826b97e","content":"采样热身报告前景、背景 IoU 与两类 mIoU；图像模型报告背景、红矩形、蓝椭圆的逐类 IoU 与三类 mIoU，再用一处边界错误解释数字的含义。","sha256":"6df7f518eaa6d443a75391daef818bd7331b266e8cc60c912cac745e6f06e95b","chapter":{"id":"course","title":"像素、边界与对象身份"},"unit":{"id":"study/pixels-instances","title":"像素、边界与对象身份"},"page":{"id":"reading-output","title":"带走的结果","number":3},"sectionTitle":"带走的结果","currentQuestion":"带走的结果","sourceSlides":[],"knowledgeBlock":null,"formulas":[],"previousBlock":"b-a90e49fc60c84e196a618b65","nextBlock":null,"prerequisites":[{"id":"b-a90e49fc60c84e196a618b65","url":"#b-a90e49fc60c84e196a618b65","role":"section-title"}],"relatedBlocks":[]},{"id":"b-e14129aa4f8e423abbda0afe","type":"h2","scope":"practice","url":"https://codingai-lec03.pages.dev/course/study/pixels-instances.html#b-e14129aa4f8e423abbda0afe","content":"## 优先实践","sha256":"0b242f211890b34b91f0dce9c1b60fa872957d318ee44d54c0ea6566dd831ab9","chapter":{"id":"course","title":"像素、边界与对象身份"},"unit":{"id":"study/pixels-instances","title":"像素、边界与对象身份"},"page":{"id":"reading-practice","title":"优先实践","number":4},"sectionTitle":"优先实践","currentQuestion":"优先实践","sourceSlides":[],"knowledgeBlock":null,"formulas":[],"previousBlock":null,"nextBlock":"b-47e55ed019bedf0e240ae160","prerequisites":[{"id":"b-47e55ed019bedf0e240ae160","url":"#b-47e55ed019bedf0e240ae160","role":"problem-context"}],"relatedBlocks":[]},{"id":"b-47e55ed019bedf0e240ae160","type":"p","scope":"practice","url":"https://codingai-lec03.pages.dev/course/study/pixels-instances.html#b-47e55ed019bedf0e240ae160","content":"可先阅读已有结果，再按问题选择是否重算。并列实验提供不同起点，不要求全部运行。","sha256":"19608cacbee0cd46460095bdebd3ee178d3479857b413a9a631e4b79f688aaef","chapter":{"id":"course","title":"像素、边界与对象身份"},"unit":{"id":"study/pixels-instances","title":"像素、边界与对象身份"},"page":{"id":"reading-practice","title":"优先实践","number":4},"sectionTitle":"优先实践","currentQuestion":"优先实践","sourceSlides":[],"knowledgeBlock":null,"formulas":[],"previousBlock":"b-e14129aa4f8e423abbda0afe","nextBlock":"b-294c54ac97501a67bb5353ac","prerequisites":[{"id":"b-e14129aa4f8e423abbda0afe","url":"#b-e14129aa4f8e423abbda0afe","role":"section-title"}],"relatedBlocks":[]},{"id":"b-294c54ac97501a67bb5353ac","type":"ul","scope":"practice","url":"https://codingai-lec03.pages.dev/course/study/pixels-instances.html#b-294c54ac97501a67bb5353ac","content":"- [LAB 11 · 从 RGB 到像素：读出、FCN 与边界](/course/notebooks/9e268706471944e3a59e6f568f9ce9a3.html)\n前景/背景/mIoU口径、RGB到mask训练、FCN与1×1读出及边界\n[在 Deepnote 打开 ↗](https://deepnote.com/workspace/ise3309jingao-5ba66b0a-4283-4f5a-ae06-3f68d3201e99/project/38755e4e-23f6-4f5f-8c68-8a2e4114b1c1/notebook/9e268706471944e3a59e6f568f9ce9a3?utm_source=openai&utm_medium=mcp&utm_campaign=openaimcp&utm_content=9e268706471944e3a59e6f568f9ce9a3&utm_term=get_notebook)","sha256":"ad954fa60070acddb3217b443b31bbcd49b21f18d1e8ff02e33fd0eb3319879d","chapter":{"id":"course","title":"像素、边界与对象身份"},"unit":{"id":"study/pixels-instances","title":"像素、边界与对象身份"},"page":{"id":"reading-practice","title":"优先实践","number":4},"sectionTitle":"优先实践","currentQuestion":"优先实践","sourceSlides":[],"knowledgeBlock":null,"formulas":[],"previousBlock":"b-47e55ed019bedf0e240ae160","nextBlock":"b-31608ee2012235246a30dc36","prerequisites":[{"id":"b-e14129aa4f8e423abbda0afe","url":"#b-e14129aa4f8e423abbda0afe","role":"section-title"},{"id":"b-47e55ed019bedf0e240ae160","url":"#b-47e55ed019bedf0e240ae160","role":"problem-context"}],"relatedBlocks":[]},{"id":"b-31608ee2012235246a30dc36","type":"li","scope":"practice","url":"https://codingai-lec03.pages.dev/course/study/pixels-instances.html#b-31608ee2012235246a30dc36","content":"- [LAB 11 · 从 RGB 到像素：读出、FCN 与边界](/course/notebooks/9e268706471944e3a59e6f568f9ce9a3.html)\n前景/背景/mIoU口径、RGB到mask训练、FCN与1×1读出及边界\n[在 Deepnote 打开 ↗](https://deepnote.com/workspace/ise3309jingao-5ba66b0a-4283-4f5a-ae06-3f68d3201e99/project/38755e4e-23f6-4f5f-8c68-8a2e4114b1c1/notebook/9e268706471944e3a59e6f568f9ce9a3?utm_source=openai&utm_medium=mcp&utm_campaign=openaimcp&utm_content=9e268706471944e3a59e6f568f9ce9a3&utm_term=get_notebook)","sha256":"ad954fa60070acddb3217b443b31bbcd49b21f18d1e8ff02e33fd0eb3319879d","chapter":{"id":"course","title":"像素、边界与对象身份"},"unit":{"id":"study/pixels-instances","title":"像素、边界与对象身份"},"page":{"id":"reading-practice","title":"优先实践","number":4},"sectionTitle":"优先实践","currentQuestion":"优先实践","sourceSlides":[],"knowledgeBlock":null,"formulas":[],"previousBlock":"b-294c54ac97501a67bb5353ac","nextBlock":"b-fbd0e43cca03697df4366716","prerequisites":[{"id":"b-e14129aa4f8e423abbda0afe","url":"#b-e14129aa4f8e423abbda0afe","role":"section-title"},{"id":"b-47e55ed019bedf0e240ae160","url":"#b-47e55ed019bedf0e240ae160","role":"problem-context"}],"relatedBlocks":[]},{"id":"b-fbd0e43cca03697df4366716","type":"p","scope":"practice","url":"https://codingai-lec03.pages.dev/course/study/pixels-instances.html#b-fbd0e43cca03697df4366716","content":"前景/背景/mIoU口径、RGB到mask训练、FCN与1×1读出及边界","sha256":"6f0b115c3c32032428a4c0d5ec4791e16d6b63aa3d643457bed758bf6af2b765","chapter":{"id":"course","title":"像素、边界与对象身份"},"unit":{"id":"study/pixels-instances","title":"像素、边界与对象身份"},"page":{"id":"reading-practice","title":"优先实践","number":4},"sectionTitle":"优先实践","currentQuestion":"优先实践","sourceSlides":[],"knowledgeBlock":null,"formulas":[],"previousBlock":"b-31608ee2012235246a30dc36","nextBlock":null,"prerequisites":[{"id":"b-e14129aa4f8e423abbda0afe","url":"#b-e14129aa4f8e423abbda0afe","role":"section-title"},{"id":"b-47e55ed019bedf0e240ae160","url":"#b-47e55ed019bedf0e240ae160","role":"problem-context"}],"relatedBlocks":[]},{"id":"b-54cff3ad2eafb279b968775b","type":"h2","scope":"readings","url":"https://codingai-lec03.pages.dev/course/study/pixels-instances.html#b-54cff3ad2eafb279b968775b","content":"## 配套讲解","sha256":"1c792c1a05326d4a07eba3b245e97c5be88fcadf7f81a2700ca1963b4c7f0299","chapter":{"id":"course","title":"像素、边界与对象身份"},"unit":{"id":"study/pixels-instances","title":"像素、边界与对象身份"},"page":{"id":"reading-readings","title":"配套讲解","number":5},"sectionTitle":"配套讲解","currentQuestion":"配套讲解","sourceSlides":[],"knowledgeBlock":null,"formulas":[],"previousBlock":null,"nextBlock":"b-506825ef2fdd72bf477f7e82","prerequisites":[{"id":"b-d74db135aa33e8baad5ff504","url":"#b-d74db135aa33e8baad5ff504","role":"problem-context"}],"relatedBlocks":[]},{"id":"b-506825ef2fdd72bf477f7e82","type":"ul","scope":"readings","url":"https://codingai-lec03.pages.dev/course/study/pixels-instances.html#b-506825ef2fdd72bf477f7e82","content":"- [CLS 02 · 从类别到区域与掩码](/course/notebooks/f9ab632ba16b4f2084669791cba6ba85.html)\n同图四输出、共享计算、FCN、RoIAlign、信息碰撞与SAM接口\n[在 Deepnote 打开 ↗](https://deepnote.com/workspace/ise3309jingao-5ba66b0a-4283-4f5a-ae06-3f68d3201e99/project/38755e4e-23f6-4f5f-8c68-8a2e4114b1c1/notebook/f9ab632ba16b4f2084669791cba6ba85?utm_source=openai&utm_medium=mcp&utm_campaign=openaimcp&utm_content=f9ab632ba16b4f2084669791cba6ba85&utm_term=get_notebook)\n- [CLS 08 · 密集预测、实例与集合输出](/course/notebooks/2d81e2a1936043349eacb71c53c0defa.html)\n像素损失、IoU口径、粗细表示、实例身份与集合输出\n[在 Deepnote 打开 ↗](https://deepnote.com/workspace/ise3309jingao-5ba66b0a-4283-4f5a-ae06-3f68d3201e99/project/38755e4e-23f6-4f5f-8c68-8a2e4114b1c1/notebook/2d81e2a1936043349eacb71c53c0defa?utm_source=openai&utm_medium=mcp&utm_campaign=openaimcp&utm_content=2d81e2a1936043349eacb71c53c0defa&utm_term=get_notebook)","sha256":"291a8d14abd034c805ac2152e5c91f4d1fa63c32680ee4d789cfcf3aa83b3907","chapter":{"id":"course","title":"像素、边界与对象身份"},"unit":{"id":"study/pixels-instances","title":"像素、边界与对象身份"},"page":{"id":"reading-readings","title":"配套讲解","number":5},"sectionTitle":"配套讲解","currentQuestion":"配套讲解","sourceSlides":[],"knowledgeBlock":null,"formulas":[],"previousBlock":"b-54cff3ad2eafb279b968775b","nextBlock":"b-5a75e3d4ca42f57cf72af22e","prerequisites":[{"id":"b-54cff3ad2eafb279b968775b","url":"#b-54cff3ad2eafb279b968775b","role":"section-title"},{"id":"b-d74db135aa33e8baad5ff504","url":"#b-d74db135aa33e8baad5ff504","role":"problem-context"}],"relatedBlocks":[]},{"id":"b-5a75e3d4ca42f57cf72af22e","type":"li","scope":"readings","url":"https://codingai-lec03.pages.dev/course/study/pixels-instances.html#b-5a75e3d4ca42f57cf72af22e","content":"- [CLS 02 · 从类别到区域与掩码](/course/notebooks/f9ab632ba16b4f2084669791cba6ba85.html)\n同图四输出、共享计算、FCN、RoIAlign、信息碰撞与SAM接口\n[在 Deepnote 打开 ↗](https://deepnote.com/workspace/ise3309jingao-5ba66b0a-4283-4f5a-ae06-3f68d3201e99/project/38755e4e-23f6-4f5f-8c68-8a2e4114b1c1/notebook/f9ab632ba16b4f2084669791cba6ba85?utm_source=openai&utm_medium=mcp&utm_campaign=openaimcp&utm_content=f9ab632ba16b4f2084669791cba6ba85&utm_term=get_notebook)","sha256":"fb96de8722fe7f9734018668e519678db20211cf892e7b05cc571067ea9c8b10","chapter":{"id":"course","title":"像素、边界与对象身份"},"unit":{"id":"study/pixels-instances","title":"像素、边界与对象身份"},"page":{"id":"reading-readings","title":"配套讲解","number":5},"sectionTitle":"配套讲解","currentQuestion":"配套讲解","sourceSlides":[],"knowledgeBlock":null,"formulas":[],"previousBlock":"b-506825ef2fdd72bf477f7e82","nextBlock":"b-d74db135aa33e8baad5ff504","prerequisites":[{"id":"b-54cff3ad2eafb279b968775b","url":"#b-54cff3ad2eafb279b968775b","role":"section-title"},{"id":"b-d74db135aa33e8baad5ff504","url":"#b-d74db135aa33e8baad5ff504","role":"problem-context"}],"relatedBlocks":[]},{"id":"b-d74db135aa33e8baad5ff504","type":"p","scope":"readings","url":"https://codingai-lec03.pages.dev/course/study/pixels-instances.html#b-d74db135aa33e8baad5ff504","content":"同图四输出、共享计算、FCN、RoIAlign、信息碰撞与SAM接口","sha256":"e0584a8ed4ddfb5b9c9abb132012919eb506480c39698c3ddd67fccd73d0a12e","chapter":{"id":"course","title":"像素、边界与对象身份"},"unit":{"id":"study/pixels-instances","title":"像素、边界与对象身份"},"page":{"id":"reading-readings","title":"配套讲解","number":5},"sectionTitle":"配套讲解","currentQuestion":"配套讲解","sourceSlides":[],"knowledgeBlock":null,"formulas":[],"previousBlock":"b-5a75e3d4ca42f57cf72af22e","nextBlock":"b-4eef245e5ad6cdb7b9492a66","prerequisites":[{"id":"b-54cff3ad2eafb279b968775b","url":"#b-54cff3ad2eafb279b968775b","role":"section-title"}],"relatedBlocks":[]},{"id":"b-4eef245e5ad6cdb7b9492a66","type":"li","scope":"readings","url":"https://codingai-lec03.pages.dev/course/study/pixels-instances.html#b-4eef245e5ad6cdb7b9492a66","content":"- [CLS 08 · 密集预测、实例与集合输出](/course/notebooks/2d81e2a1936043349eacb71c53c0defa.html)\n像素损失、IoU口径、粗细表示、实例身份与集合输出\n[在 Deepnote 打开 ↗](https://deepnote.com/workspace/ise3309jingao-5ba66b0a-4283-4f5a-ae06-3f68d3201e99/project/38755e4e-23f6-4f5f-8c68-8a2e4114b1c1/notebook/2d81e2a1936043349eacb71c53c0defa?utm_source=openai&utm_medium=mcp&utm_campaign=openaimcp&utm_content=2d81e2a1936043349eacb71c53c0defa&utm_term=get_notebook)","sha256":"13fd7efed3962e7ae1181a1c43b5745a38874692751a3ee2e455036c500ef0ff","chapter":{"id":"course","title":"像素、边界与对象身份"},"unit":{"id":"study/pixels-instances","title":"像素、边界与对象身份"},"page":{"id":"reading-readings","title":"配套讲解","number":5},"sectionTitle":"配套讲解","currentQuestion":"配套讲解","sourceSlides":[],"knowledgeBlock":null,"formulas":[],"previousBlock":"b-d74db135aa33e8baad5ff504","nextBlock":"b-d6eb1414ec3969b3d53f344e","prerequisites":[{"id":"b-54cff3ad2eafb279b968775b","url":"#b-54cff3ad2eafb279b968775b","role":"section-title"},{"id":"b-d74db135aa33e8baad5ff504","url":"#b-d74db135aa33e8baad5ff504","role":"problem-context"}],"relatedBlocks":[]},{"id":"b-d6eb1414ec3969b3d53f344e","type":"p","scope":"readings","url":"https://codingai-lec03.pages.dev/course/study/pixels-instances.html#b-d6eb1414ec3969b3d53f344e","content":"像素损失、IoU口径、粗细表示、实例身份与集合输出","sha256":"70eec56f79b131d4695a1c420bc695a02ef3863727481994de8f7ed4bd673dcc","chapter":{"id":"course","title":"像素、边界与对象身份"},"unit":{"id":"study/pixels-instances","title":"像素、边界与对象身份"},"page":{"id":"reading-readings","title":"配套讲解","number":5},"sectionTitle":"配套讲解","currentQuestion":"配套讲解","sourceSlides":[],"knowledgeBlock":null,"formulas":[],"previousBlock":"b-4eef245e5ad6cdb7b9492a66","nextBlock":null,"prerequisites":[{"id":"b-54cff3ad2eafb279b968775b","url":"#b-54cff3ad2eafb279b968775b","role":"section-title"},{"id":"b-d74db135aa33e8baad5ff504","url":"#b-d74db135aa33e8baad5ff504","role":"problem-context"}],"relatedBlocks":[]},{"id":"b-cbbb4ee3da2c7c4cc62d2df7","type":"h4","scope":"extensions","url":"https://codingai-lec03.pages.dev/course/study/pixels-instances.html#b-cbbb4ee3da2c7c4cc62d2df7","content":"#### 遇到相应问题时再展开（3 项）","sha256":"119e0bb5659d3b66061ced671f084e2b1796c470fd91613e9db2a78048731c64","chapter":{"id":"course","title":"像素、边界与对象身份"},"unit":{"id":"study/pixels-instances","title":"像素、边界与对象身份"},"page":{"id":"reading-extensions","title":"遇到相应问题时再展开（3 项）","number":6},"sectionTitle":"遇到相应问题时再展开（3 项）","currentQuestion":"遇到相应问题时再展开（3 项）","sourceSlides":[],"knowledgeBlock":null,"formulas":[],"previousBlock":null,"nextBlock":"b-b1dc84b1acfe27ad6bef2131","prerequisites":[{"id":"b-f17b5385818f75a2df4f3db4","url":"#b-f17b5385818f75a2df4f3db4","role":"problem-context"}],"relatedBlocks":[]},{"id":"b-b1dc84b1acfe27ad6bef2131","type":"ul","scope":"extensions","url":"https://codingai-lec03.pages.dev/course/study/pixels-instances.html#b-b1dc84b1acfe27ad6bef2131","content":"- [LAB 03 · 空间信息：表示、采样与掩码](/course/notebooks/49a2c030eafa4e1db6fc4f1a08c3810a.html)\n同源场景标签、插值、信息碰撞的轻量解析代码\n[在 Deepnote 打开 ↗](https://deepnote.com/workspace/ise3309jingao-5ba66b0a-4283-4f5a-ae06-3f68d3201e99/project/38755e4e-23f6-4f5f-8c68-8a2e4114b1c1/notebook/49a2c030eafa4e1db6fc4f1a08c3810a?utm_source=openai&utm_medium=mcp&utm_campaign=openaimcp&utm_content=49a2c030eafa4e1db6fc4f1a08c3810a&utm_term=get_notebook)\n- [LAB 16 · 点、框与负提示：分割规则基线](/course/notebooks/d83defa954034769a1ee94e3ea48640a.html)\n像素连通域、正负点/框提示、局部/整体多义性\n[在 Deepnote 打开 ↗](https://deepnote.com/workspace/ise3309jingao-5ba66b0a-4283-4f5a-ae06-3f68d3201e99/project/38755e4e-23f6-4f5f-8c68-8a2e4114b1c1/notebook/d83defa954034769a1ee94e3ea48640a?utm_source=openai&utm_medium=mcp&utm_campaign=openaimcp&utm_content=d83defa954034769a1ee94e3ea48640a&utm_term=get_notebook)\n- [LAB 20 · 尺度跨度与两对象匹配](/course/notebooks/9315996660f54e92847770355b648f9e.html)\n步幅占格、两对象排列、分类/位置/IoU联合匹配代价\n[在 Deepnote 打开 ↗](https://deepnote.com/workspace/ise3309jingao-5ba66b0a-4283-4f5a-ae06-3f68d3201e99/project/38755e4e-23f6-4f5f-8c68-8a2e4114b1c1/notebook/9315996660f54e92847770355b648f9e?utm_source=openai&utm_medium=mcp&utm_campaign=openaimcp&utm_content=9315996660f54e92847770355b648f9e&utm_term=get_notebook)","sha256":"e79866b4a5a13c04e04af53c771cb58efafc6484194549655394bc08f209a0cc","chapter":{"id":"course","title":"像素、边界与对象身份"},"unit":{"id":"study/pixels-instances","title":"像素、边界与对象身份"},"page":{"id":"reading-extensions","title":"遇到相应问题时再展开（3 项）","number":6},"sectionTitle":"遇到相应问题时再展开（3 项）","currentQuestion":"遇到相应问题时再展开（3 项）","sourceSlides":[],"knowledgeBlock":null,"formulas":[],"previousBlock":"b-cbbb4ee3da2c7c4cc62d2df7","nextBlock":"b-ea92cc7e755ea10d513af8e4","prerequisites":[{"id":"b-cbbb4ee3da2c7c4cc62d2df7","url":"#b-cbbb4ee3da2c7c4cc62d2df7","role":"section-title"},{"id":"b-f17b5385818f75a2df4f3db4","url":"#b-f17b5385818f75a2df4f3db4","role":"problem-context"}],"relatedBlocks":[]},{"id":"b-ea92cc7e755ea10d513af8e4","type":"li","scope":"extensions","url":"https://codingai-lec03.pages.dev/course/study/pixels-instances.html#b-ea92cc7e755ea10d513af8e4","content":"- [LAB 03 · 空间信息：表示、采样与掩码](/course/notebooks/49a2c030eafa4e1db6fc4f1a08c3810a.html)\n同源场景标签、插值、信息碰撞的轻量解析代码\n[在 Deepnote 打开 ↗](https://deepnote.com/workspace/ise3309jingao-5ba66b0a-4283-4f5a-ae06-3f68d3201e99/project/38755e4e-23f6-4f5f-8c68-8a2e4114b1c1/notebook/49a2c030eafa4e1db6fc4f1a08c3810a?utm_source=openai&utm_medium=mcp&utm_campaign=openaimcp&utm_content=49a2c030eafa4e1db6fc4f1a08c3810a&utm_term=get_notebook)","sha256":"23e6a723a12be191ae6889710e28c2341e839e8a0af1450d2c724a6d308b233c","chapter":{"id":"course","title":"像素、边界与对象身份"},"unit":{"id":"study/pixels-instances","title":"像素、边界与对象身份"},"page":{"id":"reading-extensions","title":"遇到相应问题时再展开（3 项）","number":6},"sectionTitle":"遇到相应问题时再展开（3 项）","currentQuestion":"遇到相应问题时再展开（3 项）","sourceSlides":[],"knowledgeBlock":null,"formulas":[],"previousBlock":"b-b1dc84b1acfe27ad6bef2131","nextBlock":"b-f17b5385818f75a2df4f3db4","prerequisites":[{"id":"b-cbbb4ee3da2c7c4cc62d2df7","url":"#b-cbbb4ee3da2c7c4cc62d2df7","role":"section-title"},{"id":"b-f17b5385818f75a2df4f3db4","url":"#b-f17b5385818f75a2df4f3db4","role":"problem-context"}],"relatedBlocks":[]},{"id":"b-f17b5385818f75a2df4f3db4","type":"p","scope":"extensions","url":"https://codingai-lec03.pages.dev/course/study/pixels-instances.html#b-f17b5385818f75a2df4f3db4","content":"同源场景标签、插值、信息碰撞的轻量解析代码","sha256":"93fe7c165c595944c7af5b3abd9e50016526d59fcb6d412fbfb865bd208f77c4","chapter":{"id":"course","title":"像素、边界与对象身份"},"unit":{"id":"study/pixels-instances","title":"像素、边界与对象身份"},"page":{"id":"reading-extensions","title":"遇到相应问题时再展开（3 项）","number":6},"sectionTitle":"遇到相应问题时再展开（3 项）","currentQuestion":"遇到相应问题时再展开（3 项）","sourceSlides":[],"knowledgeBlock":null,"formulas":[],"previousBlock":"b-ea92cc7e755ea10d513af8e4","nextBlock":"b-b378f89fc7d41f333daf2065","prerequisites":[{"id":"b-cbbb4ee3da2c7c4cc62d2df7","url":"#b-cbbb4ee3da2c7c4cc62d2df7","role":"section-title"}],"relatedBlocks":[]},{"id":"b-b378f89fc7d41f333daf2065","type":"li","scope":"extensions","url":"https://codingai-lec03.pages.dev/course/study/pixels-instances.html#b-b378f89fc7d41f333daf2065","content":"- [LAB 16 · 点、框与负提示：分割规则基线](/course/notebooks/d83defa954034769a1ee94e3ea48640a.html)\n像素连通域、正负点/框提示、局部/整体多义性\n[在 Deepnote 打开 ↗](https://deepnote.com/workspace/ise3309jingao-5ba66b0a-4283-4f5a-ae06-3f68d3201e99/project/38755e4e-23f6-4f5f-8c68-8a2e4114b1c1/notebook/d83defa954034769a1ee94e3ea48640a?utm_source=openai&utm_medium=mcp&utm_campaign=openaimcp&utm_content=d83defa954034769a1ee94e3ea48640a&utm_term=get_notebook)","sha256":"86302464db105c037b947be3a1cd9e3448d76128b7a7dfdd261967230ef4cd2b","chapter":{"id":"course","title":"像素、边界与对象身份"},"unit":{"id":"study/pixels-instances","title":"像素、边界与对象身份"},"page":{"id":"reading-extensions","title":"遇到相应问题时再展开（3 项）","number":6},"sectionTitle":"遇到相应问题时再展开（3 项）","currentQuestion":"遇到相应问题时再展开（3 项）","sourceSlides":[],"knowledgeBlock":null,"formulas":[],"previousBlock":"b-f17b5385818f75a2df4f3db4","nextBlock":"b-4ad6607619c35a6f8bca6059","prerequisites":[{"id":"b-cbbb4ee3da2c7c4cc62d2df7","url":"#b-cbbb4ee3da2c7c4cc62d2df7","role":"section-title"},{"id":"b-f17b5385818f75a2df4f3db4","url":"#b-f17b5385818f75a2df4f3db4","role":"problem-context"}],"relatedBlocks":[]},{"id":"b-4ad6607619c35a6f8bca6059","type":"p","scope":"extensions","url":"https://codingai-lec03.pages.dev/course/study/pixels-instances.html#b-4ad6607619c35a6f8bca6059","content":"像素连通域、正负点/框提示、局部/整体多义性","sha256":"7d1e7531a3ac0397d25516d56db48a938b043ef98467022f8eacc41563e73771","chapter":{"id":"course","title":"像素、边界与对象身份"},"unit":{"id":"study/pixels-instances","title":"像素、边界与对象身份"},"page":{"id":"reading-extensions","title":"遇到相应问题时再展开（3 项）","number":6},"sectionTitle":"遇到相应问题时再展开（3 项）","currentQuestion":"遇到相应问题时再展开（3 项）","sourceSlides":[],"knowledgeBlock":null,"formulas":[],"previousBlock":"b-b378f89fc7d41f333daf2065","nextBlock":"b-7a420a4e5aabcac8fa1961bf","prerequisites":[{"id":"b-cbbb4ee3da2c7c4cc62d2df7","url":"#b-cbbb4ee3da2c7c4cc62d2df7","role":"section-title"},{"id":"b-f17b5385818f75a2df4f3db4","url":"#b-f17b5385818f75a2df4f3db4","role":"problem-context"}],"relatedBlocks":[]},{"id":"b-7a420a4e5aabcac8fa1961bf","type":"li","scope":"extensions","url":"https://codingai-lec03.pages.dev/course/study/pixels-instances.html#b-7a420a4e5aabcac8fa1961bf","content":"- [LAB 20 · 尺度跨度与两对象匹配](/course/notebooks/9315996660f54e92847770355b648f9e.html)\n步幅占格、两对象排列、分类/位置/IoU联合匹配代价\n[在 Deepnote 打开 ↗](https://deepnote.com/workspace/ise3309jingao-5ba66b0a-4283-4f5a-ae06-3f68d3201e99/project/38755e4e-23f6-4f5f-8c68-8a2e4114b1c1/notebook/9315996660f54e92847770355b648f9e?utm_source=openai&utm_medium=mcp&utm_campaign=openaimcp&utm_content=9315996660f54e92847770355b648f9e&utm_term=get_notebook)","sha256":"fd816034dc790e92c947b11e0bc8f6df6adb8394c5bd8d086b24a71fc6a60530","chapter":{"id":"course","title":"像素、边界与对象身份"},"unit":{"id":"study/pixels-instances","title":"像素、边界与对象身份"},"page":{"id":"reading-extensions","title":"遇到相应问题时再展开（3 项）","number":6},"sectionTitle":"遇到相应问题时再展开（3 项）","currentQuestion":"遇到相应问题时再展开（3 项）","sourceSlides":[],"knowledgeBlock":null,"formulas":[],"previousBlock":"b-4ad6607619c35a6f8bca6059","nextBlock":"b-a67a92515ed378b5d611b77c","prerequisites":[{"id":"b-cbbb4ee3da2c7c4cc62d2df7","url":"#b-cbbb4ee3da2c7c4cc62d2df7","role":"section-title"},{"id":"b-f17b5385818f75a2df4f3db4","url":"#b-f17b5385818f75a2df4f3db4","role":"problem-context"}],"relatedBlocks":[]},{"id":"b-a67a92515ed378b5d611b77c","type":"p","scope":"extensions","url":"https://codingai-lec03.pages.dev/course/study/pixels-instances.html#b-a67a92515ed378b5d611b77c","content":"步幅占格、两对象排列、分类/位置/IoU联合匹配代价","sha256":"1d7535d0c02cc7bae4e9740472c8fa2e46f9a3b9067705762875bccfe9114541","chapter":{"id":"course","title":"像素、边界与对象身份"},"unit":{"id":"study/pixels-instances","title":"像素、边界与对象身份"},"page":{"id":"reading-extensions","title":"遇到相应问题时再展开（3 项）","number":6},"sectionTitle":"遇到相应问题时再展开（3 项）","currentQuestion":"遇到相应问题时再展开（3 项）","sourceSlides":[],"knowledgeBlock":null,"formulas":[],"previousBlock":"b-7a420a4e5aabcac8fa1961bf","nextBlock":null,"prerequisites":[{"id":"b-cbbb4ee3da2c7c4cc62d2df7","url":"#b-cbbb4ee3da2c7c4cc62d2df7","role":"section-title"},{"id":"b-f17b5385818f75a2df4f3db4","url":"#b-f17b5385818f75a2df4f3db4","role":"problem-context"}],"relatedBlocks":[]},{"id":"b-ee4ba5df9fa064e1db1dbba6","type":"h2","scope":"classroom","url":"https://codingai-lec03.pages.dev/course/study/pixels-instances.html#b-ee4ba5df9fa064e1db1dbba6","content":"## 接回课堂中的问题","sha256":"5214b785a21ffd090bfef76fc13e4d2fd2444ca45eccd08a1f2b475365a5634a","chapter":{"id":"course","title":"像素、边界与对象身份"},"unit":{"id":"study/pixels-instances","title":"像素、边界与对象身份"},"page":{"id":"reading-classroom","title":"接回课堂中的问题","number":7},"sectionTitle":"接回课堂中的问题","currentQuestion":"接回课堂中的问题","sourceSlides":[],"knowledgeBlock":null,"formulas":[],"previousBlock":null,"nextBlock":null,"prerequisites":[],"relatedBlocks":[]}]}