<?xml version="1.0" encoding="UTF-8"?><rss xmlns:dc="http://purl.org/dc/elements/1.1/" xmlns:content="http://purl.org/rss/1.0/modules/content/" xmlns:atom="http://www.w3.org/2005/Atom" version="2.0" xmlns:cc="http://cyber.law.harvard.edu/rss/creativeCommonsRssModule.html">
    <channel>
        <title><![CDATA[Stories by Mike Voropaev on Medium]]></title>
        <description><![CDATA[Stories by Mike Voropaev on Medium]]></description>
        <link>https://medium.com/@mikevrpv?source=rss-78f6961d289f------2</link>
        <image>
            <url>https://cdn-images-1.medium.com/fit/c/150/150/1*M--cuxKNr3bNCDu5pGzgNA.jpeg</url>
            <title>Stories by Mike Voropaev on Medium</title>
            <link>https://medium.com/@mikevrpv?source=rss-78f6961d289f------2</link>
        </image>
        <generator>Medium</generator>
        <lastBuildDate>Sat, 21 Nov 2026 00:42:25 GMT</lastBuildDate>
        <atom:link href="https://medium.com/@mikevrpv/feed" rel="self" type="application/rss+xml"/>
        <webMaster><![CDATA[yourfriends@medium.com]]></webMaster>
        <atom:link href="http://medium.superfeedr.com/" rel="hub"/>
        <item>
            <title><![CDATA[Generative 3D Texturing using AI Depth Estimation and Substance]]></title>
            <link>https://medium.com/@mikevrpv/generative-3d-texturing-using-ai-depth-estimation-and-substance-2c5503a3da9f?source=rss-78f6961d289f------2</link>
            <guid isPermaLink="false">https://medium.com/p/2c5503a3da9f</guid>
            <category><![CDATA[3d]]></category>
            <category><![CDATA[3d-texturing]]></category>
            <category><![CDATA[stable-diffusion]]></category>
            <category><![CDATA[ai]]></category>
            <category><![CDATA[texturing]]></category>
            <dc:creator><![CDATA[Mike Voropaev]]></dc:creator>
            <pubDate>Thu, 05 Jan 2023 20:45:12 GMT</pubDate>
            <atom:updated>2023-01-06T17:32:58.744Z</atom:updated>
            <content:encoded><![CDATA[<p>Generating AI wooden carved bas-relief with depth map</p><p>In this article I’m sharing my workflow of turning Midjourney bas-relief generations into 3D textures with correct height maps for displacement.</p><p>You can see more results of this workflow on my <a href="https://www.behance.net/gallery/160574951/CIVILIZATION-LOST"><strong>Behance page</strong></a>.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*9GfXTT-8Syd8NP3PwXlYZw.png" /></figure><p><strong>Prerequisites:</strong></p><ol><li>Midjourney (or any other generative AI like DALL-E or Stable Diffusion).</li><li>Stable Diffusion UI (Invoke AI and Automatic1111). You can use only AUTOMATIC1111, but I like Invoke AI UI more.</li></ol><ul><li><a href="https://github.com/AUTOMATIC1111/stable-diffusion-webui">GitHub - AUTOMATIC1111/stable-diffusion-webui: Stable Diffusion web UI</a></li><li><a href="https://github.com/invoke-ai/InvokeAI">GitHub - invoke-ai/InvokeAI: Invoke is a leading creative engine for Stable Diffusion models, empowering professionals, artists, and enthusiasts to generate and create visual media using the latest AI-driven technologies. The solution offers an industry leading WebUI, and serves as the foundation for multiple commercial products.</a></li></ul><p>3. Substance Sampler — AI algorithm used for additional Normal, Roughness, and Ambient Occlusion data. Not a must for a depth map, but it improves the quality of the final result.</p><p>4. Substance Designer — used for additional details and combining depth maps. You can use photoshop instead, but you will miss a lot of procedural texturing tools.</p><p>5. Any photo editing software to create simple masks.</p><p>6. Good GPU / CPU if you want to run it locally (I’m using 3090 GPU). Depth estimation models eat a lot of VRAM. You can also find notebooks on Google Colab to run them in the cloud.</p><p><strong>I won’t go through the installation process of those, since you can find a lot of information on YouTube.</strong></p><h3>Flowchart</h3><p>To summarize my workflow, I created this little flowchart:</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*kQJZDQrUFNMlvGw5EKhGwg.png" /><figcaption>Workflow overview</figcaption></figure><h3>Generating wood carving relief in Midjourney</h3><p>Of course, you can use Stable Diffusion for the same purpose, but I find it easier to get some fast results in Midjourney, since their new V4 model is awesome.</p><p>For this style of image I used the following text prompt:</p><blockquote>mid century wooden carved relief masterpiece showing the arrival of large plane-like two-engine spaceship on the planet with aliens, super detailed, intricate details, realistic mid century sculpture art, textures, wood cracks and paint, classic art, archeology, correct shapes, wooden panel, wooden carving, detailed, crude, by Hieronymus Bosch, by Paolo Uccello, Non-Euclidian, Paradox, in a symbolic and meaningful style</blockquote><figure><img alt="" src="https://cdn-images-1.medium.com/max/770/1*npw3sXBwI-vNsY85OKLm3w.png" /><figcaption>Midjourney result</figcaption></figure><p>As you can see, V3 image is cropped, so it makes sense to play around with prompt a bit more. But I will just fix it in Photoshop later.</p><p>Then I’m doing upscaling in Midjourney to get higher resolution of this image:</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/948/1*OO-qmT5ljCMT9bWB9OqduQ.png" /><figcaption>Upscaled Midjourney image</figcaption></figure><p>I still find some birds and other objects a bit weird and distorted, but we will fix it later.</p><h3>Fixing image with Stable Diffusion (Invoke AI UI)</h3><p>Invoke AI’s infinite canvas feature allows selective diffusion on the masked areas of the initial image:</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/917/1*873TXST1r9ptlMfvSXHSJw.png" /><figcaption>Inpainting UI in Invoke AI</figcaption></figure><p>I recommend to use specialized inpainting models:</p><ul><li><a href="https://huggingface.co/stabilityai/stable-diffusion-2-inpainting/tree/main">stabilityai/stable-diffusion-2-inpainting at main</a></li><li><a href="https://huggingface.co/runwayml/stable-diffusion-inpainting/tree/main">runwayml/stable-diffusion-inpainting at main</a></li></ul><p>To fix this weird shape on the sneaker, I’m selecting the needed part and typing a simple prompt:</p><blockquote>mid century wooden carved relief of a sneaker</blockquote><p>I’m replacing other parts of the image the same way:</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*VkYXixOy6YzCVMaja5nBKw.png" /><figcaption>Inpainting with Stable Diffusion</figcaption></figure><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*q1ZYrpVEdF9FaZAc952V3A.png" /><figcaption>Inpainting with Stable Diffusion</figcaption></figure><p>Don’t forget that you can find generated parts in Invoke AI folder on your machine. If you want a bit more consistency you can try the same seed for other inpaintings of the same objects:</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/783/1*wvMpj3V2bcodY2pJpar7ig.png" /><figcaption>Using seed</figcaption></figure><p>After all work is done, just save the merged canvas image.</p><h3>Depth map generation with AUTOMATIC1111 UI</h3><p>After that you need to make sure you have depth map script installed under “Extensions” tab:</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*oKwkmWmSHi4LZMwebOt3Kw.png" /><figcaption>Extensions in AUTOMATIC1111 UI</figcaption></figure><p>If you don’t, you can install in from “Available” tab:</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/668/1*njtIfCISN1MW_fctRGVz5g.png" /><figcaption>Available addons for AUTOMATIC1111 UI</figcaption></figure><p>You can also find the GitHub repo here:</p><p><a href="https://github.com/thygate/stable-diffusion-webui-depthmap-script">GitHub - thygate/stable-diffusion-webui-depthmap-script: High Resolution Depth Maps for Stable Diffusion WebUI</a></p><p>After installation, re-launch AUTOMATIC1111 UI and go to the “Depth” tab:</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*6zAClon9Vk2bHBYF1ereVg.png" /><figcaption>AUTOMATIC1111 depth generation</figcaption></figure><p>Here you can try different models for depth map generation. Feel free to experiment, but, personally, I’m using these three in combination:</p><p><strong>“res101”</strong> (GPU, BOOST checked), gives good equalized result:</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/520/1*16Zg8_1EcDXAT_VTNgn_YQ.png" /><figcaption>res101</figcaption></figure><p><strong>“dpt_beit_large_512 (midas 3.1)” </strong>(GPU, BOOST unchecked), nice details:</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/522/1*31_fgDNDuKk0ydd41s4qqg.png" /><figcaption>dpt_beit_large_512 (midas 3.1)</figcaption></figure><p><strong>“dpt_hybrid_384 (midas 3.0)” </strong>(GPU, BOOST checked), good in selecting main objects and accents, lacks detail:</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/521/1*l6gXgqMNdz1slWO8dsvnmw.png" /><figcaption>dpt_hybrid_384 (midas 3.0)</figcaption></figure><h3>Upscaling maps</h3><p>After generating the main artwork and it’s depth maps, I need to upscale it to 8K (texture resolution I’m working with in this example). You can use upscaling in AUTOMATIC1111 UI, or use a simple <strong>Upscayl </strong>app that has most of the popular models, but lacks parameters and some flexibility:</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*wmultrZutmOaRH3V7fBcjQ.png" /><figcaption>Upscaling with Upscayl</figcaption></figure><p>You can find Upscayl repo and read more here:</p><p><a href="https://github.com/upscayl/upscayl">GitHub - upscayl/upscayl: 🆙 Upscayl - #1 Free and Open Source AI Image Upscaler for Linux, MacOS and Windows.</a></p><h3>Substance Sampler AI: Image to texture</h3><p>To get some small details from the initial image, I’m using Substance Sampler. It can give some nice results with it’s image-to-texture algorithms, but usually I’m avoiding to create height maps here and prefer the ones I showed above.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*frRQ5T_O5eJuLtSzgpEIgA.png" /><figcaption>Substance Sampler</figcaption></figure><p>I’m generating Normal, Roughness and AO maps with the Substance Sampler.</p><h3>Substance Designer</h3><p>Substance Designer is the final tool to combine all the generated stuff into a nice PBR set of texture.</p><p>I’m combining all depth maps into one to get a nice balanced result:</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/977/1*Iij3Qnp00JYbpmRgdwSvcQ.png" /><figcaption>Combining depth maps</figcaption></figure><p>Also using a bit of blur to hide some artifacts.</p><p>Then I’m adding wood details and color to the initial image procedurally:</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/599/1*Bs2nsCMuY_MAVq3Ob8CWeA.png" /><figcaption>Substance Designer procedural texturing</figcaption></figure><p>This way I can get consistent results even with different styles and colors from Midjourney and also add some missing wood texture details.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/623/1*68oiRzyk-anbY-yezQLrVQ.png" /><figcaption>PBR texture set</figcaption></figure><h3>Final Render</h3><p>Using this workflow, I was able to get some nice textures quickly that can be applied to a decimated plane with displacement, used in game development, or even 3D printed with woodfill filaments:</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*-mLCAz0dq4f-SNkWHJ2P2w.png" /><figcaption>Final result</figcaption></figure><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/0*kKRpcCu4RuLWZrQ2" /><figcaption>Wood imitating filaments for 3D printer</figcaption></figure><p>This method can be used in procedural texturing as well as sculpting. I can imagine creating some alpha maps for ZBrush, but have not tried it yet. So I see a great potential here.</p><p>Please support me with likes and following here:</p><p><a href="https://www.behance.net/mikevrpv">Mike Voropaev on Behance</a></p><p><a href="https://www.instagram.com/mikevrpv/">https://www.instagram.com/mikevrpv</a>/</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/864/1*SMHRfe0WQS9XHxyUTvlbcg.jpeg" /></figure><p>Hope this little breakdown was helpful!</p><p>Let me know if you have any questions and stay curious!</p><img src="https://medium.com/_/stat?event=post.clientViewed&referrerSource=full_rss&postId=2c5503a3da9f" width="1" height="1" alt="">]]></content:encoded>
        </item>
    </channel>
</rss>