{"id":14871,"date":"2026-02-02T18:29:20","date_gmt":"2026-02-02T17:29:20","guid":{"rendered":"https:\/\/www.keris-studio.fr\/blog\/?p=14871"},"modified":"2026-02-03T11:42:51","modified_gmt":"2026-02-03T10:42:51","slug":"ai-flux-klein-2-inpainting","status":"publish","type":"post","link":"https:\/\/www.keris-studio.fr\/blog\/?p=14871","title":{"rendered":"AI – Flux Klein 2 – Inpainting"},"content":{"rendered":"\n

<\/p>\n\n\n

<\/a> AI – Flux Klein 2 – Inpainting<\/p>\n

<\/a>presentation<\/h1>\n

Flux 2 provides superior precision for image creation, particularly when the process involves transformation, integration, or composition. Being free and open-source, it stands as a compelling alternative to Nano banana.<\/p>\n

In architecture, certain functionalities are invaluable for rapidly transforming concepts into visual representations.<\/p>\n

Detailed explanations on Flux2 can be found here : https:\/\/docs.comfy.org\/tutorials\/flux\/flux-2-klein<\/a><\/p>\n

Unlike a standard VAE, this approach utilizes novel nodes: an Inpaint crop acting as a ‘spying glass’ and the Flux Enhancer node, which is instrumental in preventing the model from generating hallucinations.<\/p>\n

\"\"<\/a><\/p>\n

<\/a>Case Study 1 – re<\/strong>place<\/h1>\n

it uses Flux2-Klein-2-image-inpainting.json<\/em><\/p>\n

We change the road material<\/p>\n

starting image<\/p>\n

<\/p>\n

inpainting the area that will be transformed<\/p>\n

<\/p>\n

prompt<\/p>\n

<\/p>\n

Magnitude to 10<\/p>\n

<\/p>\n

result<\/p>\n

\"\"<\/a><\/p>\n<\/p>\n\n

<\/div>\n\n

it perfectly fit the geometry, light and perspective<\/p>\n

note that in that case I bypassed the second image nodes<\/p>\n

<\/p>\n

for a better result a reference can be given.<\/p>\n

I want to put the cobblestones from street 2 to street 1<\/p>\n

<\/p>\n

Prompt<\/p>\n

<\/p>\n

the prompt does all the magic. If it does not work, the prompt is bad.<\/p>\n

All links are reactivate<\/p>\n

<\/p>\n

result<\/p>\n

\"\"<\/a><\/p>\n<\/p>\n

it\u2019s a perfect blend.<\/p>\n

<\/a>Case study 2 – modify<\/h1>\n

it uses Flux2-Klein-2-image-inpainting.json<\/em><\/p>\n

This is the most complex case, we have an iconic building from Le Corbusier, we want to renovate, repaint, this is easy, but also remove the shade which is along the building. The removal process is pretty hard to achieve in a single move. Usually we should erase it and then modify the image.<\/p>\n

<\/p>\n

We give a reference, inpaint<\/p>\n

<\/p>\n

the prompt that does all<\/p>\n

<\/p>\n

here\u2019s a perfect result<\/p>\n

\"\"<\/a><\/p>\n

Effectively painting the facades with varying colors proved challenging. After numerous attempts, it was determined that the appropriate approach was not to categorize them by terms like \u00ab\u00a0main facade,\u00a0\u00bb \u00ab\u00a0side facade,\u00a0\u00bb or \u00ab\u00a0long facade.\u00a0\u00bb<\/p>\n

<\/p>\n

but west facade and north facade for instance<\/p>\n

<\/p>\n

result<\/p>\n

\"\"<\/a><\/p>\n\n

<\/div>\n\n

<\/a>Case study 3 – furnituring<\/h1>\n

it uses Flux2-Klein-inpaint-two-image-1.json<\/em><\/p>\n

There I want to add existing furniture in an empty room<\/p>\n

<\/p>\n

<\/p>\n

the sofa has to fit in the right place, right scale, right orientation<\/p>\n

prompt<\/p>\n

<\/p>\n

result<\/p>\n

\"\"<\/a><\/p>\n

Now we want to add multiple references in this room<\/p>\n

<\/p>\n

we need add nodes<\/p>\n

it\u2019s a cascading process<\/p>\n

<\/p>\n

<\/p>\n

result<\/p>\n

\"\"<\/a><\/p>\n

everything is perfectly where it should be<\/p>\n\n

<\/div>\n\n

<\/a>Case study 4 – replacing buildings<\/h1>\n

This part is harder than expected. Using Flux2-Klein-2-image-inpainting.json it is possible to place a building in an existing photograph. But depending on the focal length or composition, we can expect difficulties.<\/p>\n

site<\/p>\n

<\/p>\n

building<\/p>\n

<\/p>\n

prompt<\/p>\n

<\/p>\n

result<\/p>\n

\"\"<\/a><\/p>\n

When it is in a street and it has to perfectly match the perspective, it is a bit more tricky.<\/p>\n

For accurate integration, the characteristics of the base image must be known. Additionally, the second image should ideally share a similar perspective with the first; otherwise, the process may not function correctly.<\/p>\n

<\/p>\n

scene to work with. We want to replace the two buildings at the corner.<\/p>\n

prompt<\/p>\n

<\/p>\n

result<\/p>\n

\"\"<\/a><\/p>\n

the normalized strength has been changed<\/p>\n

<\/p>\n

Other methodology<\/p>\n

From image to 3D<\/p>\n

<\/p>\n

Sketchup Camera match<\/p>\n

<\/p>\n

import<\/p>\n

<\/p>\n

match position and camera<\/p>\n

<\/p>\n

result<\/p>\n

\"\"<\/a><\/p>\n

<\/a>Case study 5 – relighting<\/h1>\n

A global relighting could resolve the noticeable remnants of the collage in the final result.<\/p>\n

In the mask editor we cover nearly everything<\/p>\n

<\/p>\n

prompt<\/p>\n

<\/p>\n

<\/p>\n

result<\/p>\n

<\/p>\n

the image is more balanced<\/p>\n

\"\"<\/a><\/p>\n

some characters are slightly deformed (like a video game)<\/p>\n

<\/p>\n

to preserve them switch to linear<\/p>\n

<\/p>\n

back to normal<\/p>\n

<\/p>\n

\"\"<\/a><\/p>\n\n

<\/div>\n\n

Thanks to<\/p>\n

https:\/\/aistudynow.com\/flux-2-0-klein-advanced-inpainting-download-workflow\/<\/p>","protected":false},"excerpt":{"rendered":"

AI – Flux Klein 2 – Inpainting presentation Flux 2 provides superior precision for image creation, particularly when the process involves transformation, integration, or composition. Being free and open-source, it stands as a compelling alternative to Nano banana. In architecture, certain functionalities are invaluable for rapidly transforming concepts into visual representations. Detailed explanations on Flux2 … Continuer la lecture de AI – Flux Klein 2 – Inpainting<\/span> →<\/span><\/a><\/p>\n","protected":false},"author":2,"featured_media":14920,"comment_status":"open","ping_status":"open","sticky":false,"template":"","format":"standard","meta":{"footnotes":""},"categories":[190,593,25,8],"tags":[546,600,629,613,630],"class_list":["post-14871","post","type-post","status-publish","format-standard","has-post-thumbnail","hentry","category-architecture-2","category-artificial","category-etats-de-lart","category-methodologie","tag-artificial-intelligence","tag-comfyui","tag-flux2","tag-inpainting","tag-klein"],"_links":{"self":[{"href":"https:\/\/www.keris-studio.fr\/blog\/index.php?rest_route=\/wp\/v2\/posts\/14871","targetHints":{"allow":["GET"]}}],"collection":[{"href":"https:\/\/www.keris-studio.fr\/blog\/index.php?rest_route=\/wp\/v2\/posts"}],"about":[{"href":"https:\/\/www.keris-studio.fr\/blog\/index.php?rest_route=\/wp\/v2\/types\/post"}],"author":[{"embeddable":true,"href":"https:\/\/www.keris-studio.fr\/blog\/index.php?rest_route=\/wp\/v2\/users\/2"}],"replies":[{"embeddable":true,"href":"https:\/\/www.keris-studio.fr\/blog\/index.php?rest_route=%2Fwp%2Fv2%2Fcomments&post=14871"}],"version-history":[{"count":10,"href":"https:\/\/www.keris-studio.fr\/blog\/index.php?rest_route=\/wp\/v2\/posts\/14871\/revisions"}],"predecessor-version":[{"id":14939,"href":"https:\/\/www.keris-studio.fr\/blog\/index.php?rest_route=\/wp\/v2\/posts\/14871\/revisions\/14939"}],"wp:featuredmedia":[{"embeddable":true,"href":"https:\/\/www.keris-studio.fr\/blog\/index.php?rest_route=\/wp\/v2\/media\/14920"}],"wp:attachment":[{"href":"https:\/\/www.keris-studio.fr\/blog\/index.php?rest_route=%2Fwp%2Fv2%2Fmedia&parent=14871"}],"wp:term":[{"taxonomy":"category","embeddable":true,"href":"https:\/\/www.keris-studio.fr\/blog\/index.php?rest_route=%2Fwp%2Fv2%2Fcategories&post=14871"},{"taxonomy":"post_tag","embeddable":true,"href":"https:\/\/www.keris-studio.fr\/blog\/index.php?rest_route=%2Fwp%2Fv2%2Ftags&post=14871"}],"curies":[{"name":"wp","href":"https:\/\/api.w.org\/{rel}","templated":true}]}}