{"version":2,"relationshipURLBase":"https://codingai-lec03.pages.dev/course/blocks/sam-rule-baseline-explained.html","relationshipResolution":"Resolve relative relationship URLs against relationshipURLBase; complete records and absolute URLs are in blocks.","title":"LAB16：颜色连通域让提示的作用变得透明","url":"https://codingai-lec03.pages.dev/course/blocks/sam-rule-baseline-explained.html","markdown":"https://codingai-lec03.pages.dev/course/blocks/sam-rule-baseline-explained.md","pages":[{"id":"reading-opening","title":"LAB16：颜色连通域让提示的作用变得透明","unit":"提示怎样指定这一次要分的区域？","unitId":"prompted-spatial-output","chapter":"II","number":1}],"aliases":[{"oldId":"b-7970e0802e5555811cce09c3","targetId":"b-7970e0802e5555811cce09c3","page":"reading-opening","operation":"move","beforeSHA256":"d2078d5be53cd00093e26ae9649ef386e8643e28ebe6e48e561a358a89df0f78","afterSHA256":"d2078d5be53cd00093e26ae9649ef386e8643e28ebe6e48e561a358a89df0f78","reason":"Same source body in authored semantic page"},{"oldId":"b-637846ad723b545239731d0b","targetId":"b-637846ad723b545239731d0b","page":"reading-opening","operation":"move","beforeSHA256":"cc360b3357cf196b2e5ddeb14dfa0fc981c6483b9b6d354127392b41cd622396","afterSHA256":"cc360b3357cf196b2e5ddeb14dfa0fc981c6483b9b6d354127392b41cd622396","reason":"Same source body in authored semantic page"},{"oldId":"b-4c883745f010bc555e2dbe4a","targetId":"b-4c883745f010bc555e2dbe4a","page":"reading-opening","operation":"move","beforeSHA256":"0ea93c8189c66e77304ac1f4f9dde95a1d382a8f7d7323980890a7e8397e51e6","afterSHA256":"0ea93c8189c66e77304ac1f4f9dde95a1d382a8f7d7323980890a7e8397e51e6","reason":"Same source body in authored semantic page"},{"oldId":"b-b064d5bb0df98dc26d73e788","targetId":"b-b064d5bb0df98dc26d73e788","page":"reading-opening","operation":"move","beforeSHA256":"4c557624835b36a583194eddac5a7e899b4c02b7570cac64ccd7d7bb29598e5a","afterSHA256":"4c557624835b36a583194eddac5a7e899b4c02b7570cac64ccd7d7bb29598e5a","reason":"Same source body in authored semantic page"},{"oldId":"b-7d3a9e8c35bf5d6d40b74297","targetId":"b-7d3a9e8c35bf5d6d40b74297","page":"reading-opening","operation":"move","beforeSHA256":"4da8ba93b7a9d464be4d626338644c5c4d5729eb6ee6e69359bb744d7642035c","afterSHA256":"4da8ba93b7a9d464be4d626338644c5c4d5729eb6ee6e69359bb744d7642035c","reason":"Same source body in authored semantic page"},{"oldId":"b-750c91ddc8a1ad863baa481a","targetId":"b-750c91ddc8a1ad863baa481a","page":"reading-opening","operation":"move","beforeSHA256":"014d3e686af1d67cf2f7363ad92439dd1a233edd6fea947dbe71b9cdac105596","afterSHA256":"014d3e686af1d67cf2f7363ad92439dd1a233edd6fea947dbe71b9cdac105596","reason":"Same source body in authored semantic page"},{"oldId":"b-bae1440bff269aaa87ed6f79","targetId":"b-bae1440bff269aaa87ed6f79","page":"reading-opening","operation":"move","beforeSHA256":"2b9ccef7e1220e2250cceff1abd8bdd736b42f8da5a223f26b029f44b962e807","afterSHA256":"2b9ccef7e1220e2250cceff1abd8bdd736b42f8da5a223f26b029f44b962e807","reason":"Same source body in authored semantic page"}],"blocks":[{"id":"b-7970e0802e5555811cce09c3","type":"h1","scope":"sam-rule-baseline-explained","url":"https://codingai-lec03.pages.dev/course/blocks/sam-rule-baseline-explained.html#b-7970e0802e5555811cce09c3","content":"# LAB16：颜色连通域让提示的作用变得透明","sha256":"d2078d5be53cd00093e26ae9649ef386e8643e28ebe6e48e561a358a89df0f78","chapter":{"id":"II","title":"区域、像素与实例"},"unit":{"id":"prompted-spatial-output","title":"提示怎样指定这一次要分的区域？"},"page":{"id":"reading-opening","title":"LAB16：颜色连通域让提示的作用变得透明","number":1},"sectionTitle":"LAB16：颜色连通域让提示的作用变得透明","currentQuestion":"LAB16：颜色连通域让提示的作用变得透明","sourceSlides":[],"knowledgeBlock":null,"formulas":[],"previousBlock":null,"nextBlock":"b-637846ad723b545239731d0b","prerequisites":[{"id":"b-637846ad723b545239731d0b","url":"#b-637846ad723b545239731d0b","role":"problem-context"}],"relatedBlocks":[]},{"id":"b-637846ad723b545239731d0b","type":"p","scope":"sam-rule-baseline-explained","url":"https://codingai-lec03.pages.dev/course/blocks/sam-rule-baseline-explained.html#b-637846ad723b545239731d0b","content":"点和框指定本次目标，但一个点仍可能同时指向局部与整体。SAM 的可复用图像表示和轻量提示解码使交互更方便；当前 LAB16 是透明规则对照，应准确区分规则输出、教学标注和真正模型输出。","sha256":"cc360b3357cf196b2e5ddeb14dfa0fc981c6483b9b6d354127392b41cd622396","chapter":{"id":"II","title":"区域、像素与实例"},"unit":{"id":"prompted-spatial-output","title":"提示怎样指定这一次要分的区域？"},"page":{"id":"reading-opening","title":"LAB16：颜色连通域让提示的作用变得透明","number":1},"sectionTitle":"LAB16：颜色连通域让提示的作用变得透明","currentQuestion":"LAB16：颜色连通域让提示的作用变得透明","sourceSlides":[],"knowledgeBlock":null,"formulas":[],"previousBlock":"b-7970e0802e5555811cce09c3","nextBlock":"b-4c883745f010bc555e2dbe4a","prerequisites":[{"id":"b-7970e0802e5555811cce09c3","url":"#b-7970e0802e5555811cce09c3","role":"section-title"}],"relatedBlocks":[]},{"id":"b-4c883745f010bc555e2dbe4a","type":"p","scope":"sam-rule-baseline-explained","url":"https://codingai-lec03.pages.dev/course/blocks/sam-rule-baseline-explained.html#b-4c883745f010bc555e2dbe4a","content":"当前 LAB16 输入 48×64 RGB 图像，含红色外壳、深红内部零件与蓝物体。代码从正点像素取颜色，选择颜色距离在容差内的像素，再保留与正点四邻连通的部分。负点排除与负点颜色太相似的位置；框只限制允许区域。","sha256":"0ea93c8189c66e77304ac1f4f9dde95a1d382a8f7d7323980890a7e8397e51e6","chapter":{"id":"II","title":"区域、像素与实例"},"unit":{"id":"prompted-spatial-output","title":"提示怎样指定这一次要分的区域？"},"page":{"id":"reading-opening","title":"LAB16：颜色连通域让提示的作用变得透明","number":1},"sectionTitle":"LAB16：颜色连通域让提示的作用变得透明","currentQuestion":"LAB16：颜色连通域让提示的作用变得透明","sourceSlides":[],"knowledgeBlock":null,"formulas":[],"previousBlock":"b-637846ad723b545239731d0b","nextBlock":"b-b064d5bb0df98dc26d73e788","prerequisites":[{"id":"b-7970e0802e5555811cce09c3","url":"#b-7970e0802e5555811cce09c3","role":"section-title"},{"id":"b-637846ad723b545239731d0b","url":"#b-637846ad723b545239731d0b","role":"problem-context"}],"relatedBlocks":[]},{"id":"b-b064d5bb0df98dc26d73e788","type":"p","scope":"sam-rule-baseline-explained","url":"https://codingai-lec03.pages.dev/course/blocks/sam-rule-baseline-explained.html#b-b064d5bb0df98dc26d73e788","content":"已有本地输出的面积为：紧容差 110，宽容差 837，增加外壳负点后 110，加框后 156。紧容差选择内部零件，宽容差把整个外壳纳入；框的面积 156 也可能截断物体。这些变化能直接从规则解释。","sha256":"4c557624835b36a583194eddac5a7e899b4c02b7570cac64ccd7d7bb29598e5a","chapter":{"id":"II","title":"区域、像素与实例"},"unit":{"id":"prompted-spatial-output","title":"提示怎样指定这一次要分的区域？"},"page":{"id":"reading-opening","title":"LAB16：颜色连通域让提示的作用变得透明","number":1},"sectionTitle":"LAB16：颜色连通域让提示的作用变得透明","currentQuestion":"LAB16：颜色连通域让提示的作用变得透明","sourceSlides":[],"knowledgeBlock":null,"formulas":[],"previousBlock":"b-4c883745f010bc555e2dbe4a","nextBlock":"b-7d3a9e8c35bf5d6d40b74297","prerequisites":[{"id":"b-7970e0802e5555811cce09c3","url":"#b-7970e0802e5555811cce09c3","role":"section-title"},{"id":"b-637846ad723b545239731d0b","url":"#b-637846ad723b545239731d0b","role":"problem-context"}],"relatedBlocks":[]},{"id":"b-7d3a9e8c35bf5d6d40b74297","type":"p","scope":"sam-rule-baseline-explained","url":"https://codingai-lec03.pages.dev/course/blocks/sam-rule-baseline-explained.html#b-7d3a9e8c35bf5d6d40b74297","content":"因此它适合回答“提示和算法假设共同怎样改变结果”，不能回答“SAM 学会了什么”。它确实读入像素并输出 mask，但没有可学习参数，没有加载 SAM checkpoint，也没有语义泛化证据。保留这项透明练习，比给它错误模型身份更有教学价值。","sha256":"4da8ba93b7a9d464be4d626338644c5c4d5729eb6ee6e69359bb744d7642035c","chapter":{"id":"II","title":"区域、像素与实例"},"unit":{"id":"prompted-spatial-output","title":"提示怎样指定这一次要分的区域？"},"page":{"id":"reading-opening","title":"LAB16：颜色连通域让提示的作用变得透明","number":1},"sectionTitle":"LAB16：颜色连通域让提示的作用变得透明","currentQuestion":"LAB16：颜色连通域让提示的作用变得透明","sourceSlides":[],"knowledgeBlock":null,"formulas":[],"previousBlock":"b-b064d5bb0df98dc26d73e788","nextBlock":"b-750c91ddc8a1ad863baa481a","prerequisites":[{"id":"b-7970e0802e5555811cce09c3","url":"#b-7970e0802e5555811cce09c3","role":"section-title"},{"id":"b-637846ad723b545239731d0b","url":"#b-637846ad723b545239731d0b","role":"problem-context"}],"relatedBlocks":[]},{"id":"b-750c91ddc8a1ad863baa481a","type":"p","scope":"sam-rule-baseline-explained","url":"https://codingai-lec03.pages.dev/course/blocks/sam-rule-baseline-explained.html#b-750c91ddc8a1ad863baa481a","content":"读完可以把同一个正点移到蓝物体，预测容差和负点的作用；或对照官方 SAM predictor 的输入输出接口，标明哪些步骤仍需真实权重推理。此处不把未来模型运行写成已经完成。","sha256":"014d3e686af1d67cf2f7363ad92439dd1a233edd6fea947dbe71b9cdac105596","chapter":{"id":"II","title":"区域、像素与实例"},"unit":{"id":"prompted-spatial-output","title":"提示怎样指定这一次要分的区域？"},"page":{"id":"reading-opening","title":"LAB16：颜色连通域让提示的作用变得透明","number":1},"sectionTitle":"LAB16：颜色连通域让提示的作用变得透明","currentQuestion":"LAB16：颜色连通域让提示的作用变得透明","sourceSlides":[],"knowledgeBlock":null,"formulas":[],"previousBlock":"b-7d3a9e8c35bf5d6d40b74297","nextBlock":"b-bae1440bff269aaa87ed6f79","prerequisites":[{"id":"b-7970e0802e5555811cce09c3","url":"#b-7970e0802e5555811cce09c3","role":"section-title"},{"id":"b-637846ad723b545239731d0b","url":"#b-637846ad723b545239731d0b","role":"problem-context"}],"relatedBlocks":[]},{"id":"b-bae1440bff269aaa87ed6f79","type":"h2","scope":"sam-rule-baseline-explained","url":"https://codingai-lec03.pages.dev/course/blocks/sam-rule-baseline-explained.html#b-bae1440bff269aaa87ed6f79","content":"## 继续阅读与实验","sha256":"2b9ccef7e1220e2250cceff1abd8bdd736b42f8da5a223f26b029f44b962e807","chapter":{"id":"II","title":"区域、像素与实例"},"unit":{"id":"prompted-spatial-output","title":"提示怎样指定这一次要分的区域？"},"page":{"id":"reading-opening","title":"LAB16：颜色连通域让提示的作用变得透明","number":1},"sectionTitle":"LAB16：颜色连通域让提示的作用变得透明","currentQuestion":"LAB16：颜色连通域让提示的作用变得透明","sourceSlides":[],"knowledgeBlock":null,"formulas":[],"previousBlock":"b-750c91ddc8a1ad863baa481a","nextBlock":null,"prerequisites":[{"id":"b-7970e0802e5555811cce09c3","url":"#b-7970e0802e5555811cce09c3","role":"section-title"},{"id":"b-637846ad723b545239731d0b","url":"#b-637846ad723b545239731d0b","role":"problem-context"}],"relatedBlocks":[]}]}