

Create a clean, modern scientific illustration of two Multiple Instance Learning (MIL) scenarios, split into two panels. Left panel: Drug Discovery. - Positive Bag: A light blue background with a central molecule (e.g., a kinase inhibitor) surrounded by multiple conformations, one highlighted in gold (active). Label: "Positive: at least one active". - Negative Bag: A light gray background with a **structurally different molecule** (e.g., an unrelated scaffold, not a conformation of the positive one) surrounded by all gray, inactive conformations. Label: "Negative: all inactive". Right panel: Cat Image Classification. - Positive Bag: A grid-split image with patches containing a cat highlighted in green. Label: "Positive: at least one target patch". - Negative Bag: A grid-split image with all gray patches (no cat). Label: "Negative: all background". Style: Clean lines, soft colors, scientific illustration, modern, non-photorealistic, consistent with the reference image layout.
描述你想生成的图片信息,如:画面里有什么风格/元素/色彩
Shift+回车 = 换行
AI可能会犯错误或不完全遵循,请谨慎使用。
同款 0
评论 0