超美的青花瓷的星奇摇!
#Prompt
#BlueAndWhitePorcelain #ChineseClassicalDance #ChineseStyleDance
利用H3 生成
<Subject 1>: Reference Image < Image 1 > – Target Character. Strictly adhere to the following character specifications: dark brown hair styled in an updo (with braids); fair skin; delicate facial features with a gentle smile; a blue-and-white, semi-sheer lace/tulle haute couture gown (featuring a *qinghua*—blue-and-white porcelain—aesthetic). The gown includes a high, Chinese-style mandarin collar, a sheer bodice, 3D blue floral embroidery, cascading ruffles, a high-slit floor-length skirt, and an open-back design with crisscross lacing. Accessories include blue-and-white floral hair ornaments with tassels, dangling tassel earrings, and silver strappy high-heeled sandals. Ensure a consistent character design (three-view style) with stable, well-defined facial features. < Image 1 > serves solely as a reference for the character’s appearance and identity; it is not to be used as the first frame of the target video.
<Subject 2>: Background Environment. Adopt the background space, ambient lighting, color palette, and depth-of-field atmosphere from < Video 1 >. Retain only the background and lighting; do not include the character present in < Video 1 >.
<Subject 3>: Visual Style. Defined by Reference Image < Image 1 >. The overall aesthetic should be realistic, resembling footage captured on an iPhone.
<Video 1>: Source Video. This serves as the definitive and exclusive reference for all actions, poses, limb movements, body rotations, pacing, pauses, camera movements, cuts, zooms, and compositional perspective. Do not alter any of the actions.
Summary: [Motion Transfer + Reference-Based Generation] The target video is created by transferring the motion from < Video 1 > onto the character defined in < Subject 1 >, while applying the visual style specified in < Subject 3 >. **Replicate 100% of the actions and camera movements from the source video; do not add, remove, advance, or delay any physical movements.**
🤖 Made with AI
黑白相生转场 !
年轻时总爱分对错、论黑白,
觉得世界非此即彼。
成熟以后才懂,黑白本就相生。
成年人的世界哪有那么多绝对,
大多数时候,我们都是在灰色地带里找平衡。
你是从哪件事开始,不再执着于“非黑即白”的?
#prompt #BlackAndWhite #RightAndWrong #OnTheEdgeOfGrey #Black&White #Right&Wrong #GreyEdge #MoralAmbiguity #GreyZone #BoundaryOfGoodAndEvil
利用H 3 动作迁移
[Motion Transfer + Reference-Based Generation] The target video is created by transferring the motion from < Video 1 > onto the character defined in < Subject 1 >, while applying the visual style specified in < Subject 3 >. **Replicate 100% of the actions and camera movements from the source video; do not add, remove, advance, or delay any physical movements.**
9
29
514
早啊!又是阳光灿烂的一天。
每天都是新的开始
马尔福大小姐
问你吃饭没有?
#马尔福大小姐 #哈利波特
#DracoMalfoyGenderbend #HarryPotterAU #MalfoyLady #HPfanart #Malfoy
🤖 Made with AI
5
10
351
现在我也是
马尔福大小姐的忠实粉丝。
#prompt
photorealistic portrait photography, 2x2 contact sheet, four-panel collage, same adult woman in all four frames, consistent facial identity, long champagne blonde hair, Korean beauty aesthetic, natural pink blush, subtle eyeliner, black oversized knit cardigan, white collared shirt, dark emerald green striped tie, gold hoop earrings, cozy indoor setting, beige curtains, sheer white curtains, white armchair, soft natural window lighting, iPhone candid photography, realistic skin pores, detailed hair strands, natural facial expressions, authentic smartphone camera rendering, warm neutral color grading, soft shadows, high detail, realistic photography, 1:1 square composition.
Negative prompt: anime, illustration, CGI, 3D render, plastic skin, excessive skin smoothing, artificial facial features, inconsistent identity, different outfits, extra fingers, malformed hands, distorted face, duplicated person, text, watermark, logo.
#马尔福大小姐 #哈利波特
🤖 Made with AI










