Virtual idol creative video
Virtual idol creative video
虚拟偶像视频创作：C4D/AE/剪映
Virtual idol creative video Tools used: C4D/AE/Clipping
Various AI generated videos
各种AI生成视频创作：RUNWAY/SD/MJ/COMFYUI/豆包
Virtual idol creative video Tools used: C4D/AE/剪映/
AR虚拟现实
这是一个结合了AR眼镜产品与巨量引擎文旅的短视频。从AR眼镜的第一视角玩“羊了个羊”游戏到以各种虚拟方式体验各地景点旅游，即展现未来科技足不出户的出行方式，又按客户需求介绍了北京欢乐谷、鼓浪屿、英国、澳门科技馆的命题。
This is a short video that combines AR glasses products with a massive engine for cultural tourism. From the first perspective of AR glasses playing the game "Sheep Takes a Sheep" to experiencing various scenic spots and tourism in various virtual ways, it not only showcases the future technology of staying at home, but also introduces the themes of Beijing Happy Valley, Gulangyu, UK, and Macau Science and Technology Museum according to customer needs.
2D&3D混合现实
3D与2D次元的混合世界，用这种一镜到底动态式插画与3D虚拟人混搭的方式诠释杭州citywalk独特的景观之美，使用AE技术与手绘插画与3D渲染合成完成
The hybrid world of 3D and 3D dimensions, using this one shot to the end dynamic illustration and 3D virtual human hybrid approach to interpret the unique landscape beauty of Hangzhou CityWalk, not only reflects the empowerment of sports with AR glasses, but also meets the needs of the massive engine of cultural tourism.
混合现实
这是一个融合了实拍景点视频的AR眼镜功能介绍。介绍AR技术的随身导游功能，借助sd生成的素材与AE的动画处理技术和AI处理图像技术共通协作而成。
This is an introduction to the AR glasses function that integrates real-life scenic spot videos. Introduce the portable tour guide function of AR technology, introduce the historical background of the scenic spot, and choose to open it in various parallel universe styles. With the development of technology, AR glasses will combine VR's immersive virtual world experience with augmented reality capabilities.
进入元宇宙
这是一个融合了实拍景点视频的AR眼镜功能介绍。介绍AR技术的随身导游功能，借助sd生成的素材与AE的动画处理技术和AI处理图像技术共通协作而成。
Displaying the regional characteristics of the scenic spots according to customer requirements: from Chimelong Super Mario Zoo to the ancient and fragrant Shaoxing Water Town, and then to the modern Hangzhou City, the concept of "metaverse" that embraces diversity and inclusiveness is reflected through silky transitions and clever logical transitions.
虚拟人创意视频
这是一个结合了虚拟人宣传的唯美主义创意视频。主要通过C4D建模、动作、渲染等技术和AE后期来完成的。
This is a short video that combines AR glasses products with a massive engine for cultural tourism. From taking real photos wearing AR glasses to experiencing various tourist attractions from the first person perspective after wearing glasses
The latest
progress in AI technology mining
AI技术挖掘最新进展
通过AI生成电商设计，让文字与主题表达更加和谐！
通过掌握书写关键词的技巧，可以做到电商设计分分钟出图，关键词技巧包括：场景按照产品特质来！通过营造出的场景，最大化的体现产品的特色，通过提示词的恰当表述来生成更佳贴近主题感受的文字风格！
By mastering the techniques of writing key words, you can produce e-commerce designs in minutes. The key word techniques include: scenes tailored to product characteristics! By creating scenes that maximize the display of product features and using appropriately phrased prompts, you can generate text styles that better align with the thematic tone!
AI生成电商大促活动系列KV，完美还原产品外观，取代3D软件制作。产品KV包括：详情页首图、A+、海报、banner。保证产品外观的还原度，不再使用C4D\blende\3D MAX等软件，可节省大量劳动时间和精力～
AI generated e-commerce promotion event series KV, perfectly restoring product appearance and replacing 3D software production. The product KV includes: details page header image A+、 poster banner。 Ensure the restoration of product appearance and no longer use software such as C4D \ blende \ 3D MAX, which can save a lot of labor time and energy
电商大促活动系列KV快速生成，完美还原产品外观，取代3D软件制作
用鼠标向右滑动👉
拟定好文字和用途，提供主图，把如何排版交给AI的尝试
下面是通过尝试得出的生成结果，拟定好文字与用途，仅仅提供给AI主图，把怎样排版交给AI，目前尝试让AI生成杂志封面排版、书籍封面排版、海报排版、儿童绘本排版、产品包装、VI设计。
By training the model, it is possible to generate works with relatively fewer bugs, consistent character features, and consistent prop element styles! Usually, this type of style is niche or labeled as difficult to find similar works on the internet to meet the specific needs of commercial activities!
套用排版方式，更改主体和文字的尝试
因为每种视觉设计都似乎有自己的排版套路，比如电商版式、报纸排版、海报排版、抖音封面排版、通知排版、节日推送页排版，我们可以更改文字信息以及主体，就可得到一张宸人心意的视觉设计！
Since each type of visual design seems to have its own layout style, such as e-commerce layouts, newspaper layouts, poster layouts, TikTok cover layouts, notification layouts, and holiday push page layouts, we can modify the text and main elements to create a visually appealing design that resonates with the audience!
高清打印需求：超高清工作流，颗粒度级别的呈现！
通过comfyui照片颗粒度级别的工作流搭建，能够生成指定参考照片、文生图、或清晰化处理的工作流，此工作流多用于摄影照片以及写实风格，它不同于普通的高清放大功能，而是可以生成质感光影非常丰富的细节，非常适用于杂志封面、户外海报、产品细节图的展示！此工作流生成尺寸目前可满足5米X7米的户外海报。
By building a workflow at the granularity level of Comfyui photos, it is possible to generate specified reference photos, cultural and biological images, or clear processing workflows. This workflow is mostly used for photography photos and realistic styles. It is different from the ordinary high-definition zoom function, but can generate details with rich texture and light and shadow, which is very suitable for displaying magazine covers, outdoor posters, and product detail images! This workflow can currently generate outdoor posters with a size of 5 meters by 7 meters.
一人创作并运营IP品牌公众号不再是梦
通过特定形象lora模型训练，我们可以零成本打造自己的IP形象，并以IP形象为主导创作搞笑、治愈、温馨的条漫，它可以是干货的安利，也可以是料理教程或技能教程等，总之，条漫的形式不仅可以吸引到中年打工族群体还可以吸引到学生青少年群体，总之，条漫的阅读形式是内卷压力时代的一剂良药！也是个人IP打造的最有效方法！
Through AI, we can create our own IP image at zero cost and create humorous, healing, and heartwarming comics based on the IP image. It can be a way to sell practical knowledge, cooking tutorials, or skill tutorials. In short, the form of comics can not only attract middle-aged working people but also students and teenagers. In short, the reading form of comics is a good medicine for the era of internal pressure! It is also the most effective way to create a personal IP!
用鼠标向右滑动👉
lora模型训练：提升特殊风格模型泛化处理力，对语意正确反馈！
通过lora模型训练，能生成bug相对较少的人物特征一致性和道具元素风格一致性的作品！通常这类风格是小众化风格，或者被标签为在网络上很难寻找到类似的作品从而满足商业活动的特定需求！
By training the lora model, it is possible to generate works with relatively fewer bugs, consistent character features, and consistent prop element styles! Usually, this type of style is niche or labeled as difficult to find similar works on the internet to meet the specific needs of commercial activities!
连载漫画：保证主角与配角的形象稳定性和环境同一性。
通过模型训练，能生成bug相对较少的人物特征一致性和道具元素风格一致性的作品！通常这类风格是小众化风格，或者被标签为在网络上很难寻找到类似的作品从而满足商业活动的特定需求！
By training the model, it is possible to generate works with relatively fewer bugs, consistent character features, and consistent prop element styles! Usually, this type of style is niche or labeled as difficult to find similar works on the internet to meet the specific needs of commercial activities!
从主角设定——三视图——关于他的连载故事
通过算法，我们可以创造自己的人物形象以及归属他的宇宙世界观，我们可以把那些文字小说变成连载漫画故事，把通过文字想象的人物形象用AI生成出来，又或者通过连载漫画故事进一步生成视频，发布到各大平台上，让更多的人看到作者的精彩故事小说！
Through algorithms, we can create our own character designs and their corresponding universe worldview. We can transform written novels into serialized comic stories, generate AI visuals of characters imagined through text, or further develop these serialized comics into videos, publishing them on major platforms to showcase the author's captivating story novels to a wider audience!
用鼠标向右滑动👉
通过lora微调技术让算法去拟合，更多还原风格的细腻度。
通过模型训练，能生成bug相对较少的人物特征一致性和道具元素风格一致性的作品！通常这类风格是小众化风格，或者被标签为在网络上很难寻找到类似的作品从而满足商业活动的特定需求！
Through model training, it is possible to generate works with relatively fewer bugs, ensuring character consistency and a unified style in props! Typically, such styles are niche or labeled as hard to find similar works online, thus meeting the specific needs of commercial activities!
用鼠标向右滑动👉
