Stable Diffusion - 编辑生成 (OpenPose Editor) 相同人物姿势的图像

欢迎关注我的CSDN:https://spike.blog.csdn.net/

本文地址:https://spike.blog.csdn.net/article/details/131818943

OpenPose Editor 是 Stable Diffusion 的扩展插件,可以自定义人物的姿势和表情,以及生成深度、法线和边缘图等信息,打造个性化的人物绘画,或者修复和提高图像的质量,是 ControlNet 中 OpenPose 的有力补充,即 sd-webui-openpose-editor

1. 图像反推提示词

模型:DreamShaper 7

使用 图生图 中 反推图像提示词 功能,融合 CLIP 和 DeepBooru,再添加质量关键词,获得与原图风格相同的提示词组,即:

a woman sitting on a boat in the ocean wearing a hat and a white dress with a slit down the side,Daphne Fedarb,olive skin,a bronze sculpture,aestheticism,1girl,blonde hair,boat,day,dress,full body,hat,horizon,lips,long hair,looking at viewer,ocean,outdoors,pier,pool,poolside,railing,realistic,sitting,sky,sleeveless,sleeveless dress,smile,solo,stairs,sun hat,water,watercraft,white dress,lora:ClothingAdjuster2:-0.4,

默认负面提示词:

(BadDream,UnrealisticDream,badhandv4,bad_prompt_version2-neg,EasyNegative,ng_deepnegative_v1_75t:0.9),(worst quality, low quality:1.3),(depth of field, blurry:1.2),(greyscale, monochrome:1.1),croped,lowres,text,jpeg artifacts,(logo,signature,watermark,username,artist name,title:1.3),

2. 生成与编辑姿势

使用 ControlNet 的 OpenPose 提取人物姿势,因遮挡原因,部分骨骼点缺失,即:

使用编辑人物姿势,增加左手关键点,完成之后,发送至 ControlNet。支持功能:增加缺失的骨骼点、调整姿势比例,调整画布大小等。即:

插件:sd-webui-openpose-editor

其中,ControlNet 配置选择:缩放后填充空白预处理结果作为输入(Preview as Input)

其他配置:

3. 图像效果

图像信息:

a woman sitting on a boat in the ocean wearing a hat and a white dress with a slit down the side,Daphne Fedarb,olive skin,a bronze sculpture,aestheticism,1girl,blonde hair,boat,day,dress,full body,hat,horizon,lips,long hair,looking at viewer,ocean,outdoors,pier,pool,poolside,railing,realistic,sitting,sky,sleeveless,sleeveless dress,smile,solo,stairs,sun hat,water,watercraft,white dress,lora:ClothingAdjuster2:-0.4,

Negative prompt: (BadDream,UnrealisticDream,badhandv4,bad_prompt_version2-neg,EasyNegative,ng_deepnegative_v1_75t:0.9),(worst quality, low quality:1.3),(depth of field, blurry:1.2),(greyscale, monochrome:1.1),croped,lowres,text,jpeg artifacts,(logo,signature,watermark,username,artist name,title:1.3),

Steps: 35, Sampler: DDIM, CFG scale: 7, Seed: 2382679957, Face restoration: CodeFormer, Size: 768x1024, Model hash: ed989d673d, Model: Dreamshaper_7, Denoising strength: 0.3, Clip skip: 2, ADetailer model: face_yolov8n.pt, ADetailer prompt: "detailed face,", ADetailer confidence: 0.3, ADetailer dilate/erode: 4, ADetailer mask blur: 4, ADetailer denoising strength: 0.4, ADetailer inpaint only masked: True, ADetailer inpaint padding: 32, ADetailer version: 23.7.6, ControlNet 0: "preprocessor: none, model: control_v11p_sd15_openpose [cab727d4], weight: 1, starting/ending: (0, 1), resize mode: Resize and Fill, pixel perfect: True, control mode: Balanced, preprocessor params: (512, -1, -1)", Hires upscale: 2, Hires steps: 20, Hires upscaler: 4x-UltraSharp, Lora hashes: "ClothingAdjuster2: f038e3a5b67b", Version: v1.4.0

所生成的图像与原图,姿势保持相同,同时,增加左手,效果如下:

  • PS 合并图像:新建、移动、自由变换(Command T) 等比例修改图像大小。

参考:

相关推荐
weiwei2284418 天前
文生图模型Stable Diffusion使用详解
stable diffusion·文生图
Hello123网站18 天前
DreamStudio-Stable Diffusion官方绘画工具
stable diffusion·ai工具
如若12318 天前
CompVis Stable Diffusion 正确安装与运行指南
stable diffusion
全息数据24 天前
DDPM代码讲解【详细!!!】
深度学习·stable diffusion·多模态·ddpm
UWA1 个月前
虚拟相机的最佳实践参考是什么
数码相机·editor·memory·rendering
老鱼说AI1 个月前
当自回归模型遇上扩散模型:下一代序列预测模型详解与Pytorch实现
人工智能·pytorch·深度学习·神经网络·语言模型·自然语言处理·stable diffusion
我希望的一路生花1 个月前
Nik Collection 6.2全新版Nik降噪锐化调色PS/LR插件
人工智能·计算机视觉·设计模式·stable diffusion·aigc
GetcharZp1 个月前
玩转AI绘画,你只差一个节点式“魔法”工具——ComfyUI 保姆级入门指南
人工智能·stable diffusion
Seeklike1 个月前
diffuxers学习--AutoPipeline
人工智能·python·stable diffusion·diffusers
游戏AI研究所1 个月前
ComfyUI 里的 Prompt 插值器(prompt interpolation / text encoder 插值方式)的含义和作用!
人工智能·游戏·机器学习·stable diffusion·prompt·aigc