Umkhiqizi we-AI Depth Map
Ukusetshenziswa kwebhizinisi OK
Amamodeli angama-380+
Akukho phawu lwamanzi
Akukho ubhaliso olungenayo
Imodeli:
+ GPT-5, Claude, Gemini
Dala imaphu yobubanzi kusuka kunoma iyiphi isithombe. Isetshenziswa njenge-ControlNet conditioning input, ye-parallax / 3D effects, noma njengendawo yokuqalisa ye-AR/VR scene reconstruction. Isebenza nge-Marigold (emuva, ecacile) noma i-fal Depth (ekhawulelwe).
Izinketho ezithuthukisiwe
Imiphumela
Ama-token aphansi.
Thola ama-token amaningi
❤️ Uthanda i-Free.ai? Ngisho nabahlobo bakho!
Ngena ukuze uthole i-link yokuxhumanisa futhi uthole ama-token angama-25,000 ngamunye umngane.
Ufuna okuningi?
Bhala mahhala nge-30K tokens / ngosuku + 10K ibhonasi
Bhala
Kuqedwa isicelo sakho...
Dala imaphu yobubanzi kusuka kunoma iyiphi isithombe nge-AI. I-Marigold emahhala (eqinile, egcinwe ngokwayo) noma ububanzi obusheshayo obuphezulu — ikhipha ububanzi be-ControlNet-elungele i-PNG ye-parallax, AR/VR, ukuvuselelwa kwesithombe se-3D.
Indlela Yokusetshenziswa Umkhiqizi we-AI Depth Map
1
Faka umngenisi wakho
Bhala umbhalo, ulayishe ifayela, noma uchaza ofuna ukukwenza. Akukho akhawunti edingekayo.
2
Chofoza ukwakha
I-AI yethu isebenza isicelo sakho emaminithini usebenzisa amamodeli avulekile angcono kakhulu.
3
Layisha phezulu futhi uhlukanise
Layisha phezulu, kopela, noma uhlukanise imiphumela yakho. Imahhala ukusetshenziswa komuntu siqu nokwebhizinisi.
Sebenzisa lesi sisetshenziswa nge-API
Yenza lokhu sixhobo sisebenze ngokuzenzakalela kusuka kukhodi yakho. I-REST ehambisanayo ne-OpenAI, ukuqinisekisa i-token ye-bearer, akukho SDK edingekayo. Izindleko ze-token zifana ne-web interface.
curl -X POST https://api.free.ai/v1/image/generate/ \
-H "Authorization: Bearer sk-free-..." \
-H "Content-Type: application/json" \
-d '{"prompt": "A sunset over mountains", "model": "sdxl"}'Umkhiqizi we-AI Depth Map — FAQ
Layisha phezulu noma iyiphi isithombe, uthole imaphu yobubanzi emuva — isithombe esibomvu lapho i-pixel ngayinye ikhoda khona ukuthi ibanga liphi isikhundla esisuka kwikhamera. Sebenzisa njenge-ControlNet input yesithombe esinobubanzi obuqinile, njenge-depth pass ye-paralax / 3D effects ku-After Effects noma ku-Blender, noma njengendawo yokuqalisa ye-AR / VR scene reconstruction. Amamodeli amabili: i-Marigold emahhala (eqinile, ehamba kancane) noma ububanzi obukhawulelwe (iminingwane encane, ihamba ngokushesha ku-inputs enkulu).
Yebo. I-Marigold isebenza kuma-A100 GPUs ethu asebenza ngokuzimela - isuka epulazini lakho le-token yansuku zonke (2,500 / ngosuku olungaziwayo, 10,000 / ngosuku olubhalisiwe). I-map ngayinye yobubanzi i-token ye-1,500. I-premium fast-depth ibiza i-baseline efanayo kodwa ingakala kahle kuma-inputs we-4K +.
I-Marigold (iphutha) inikeza ububanzi obuqinile, obuyiqiniso kakhulu — i-fine-tuned kusuka ku-Stable Diffusion, ithatha ububanzi obuncane emithini / izinwele / izimbobo zendwangu. I-Premium ishesha kakhulu kuma-inputs we->2K futhi ithembekile kakhulu kuma-wide-angle shots. Ukulawulwa kwe-Net conditioning noma ukuvuselelwa kwe-3D, iphutha ku-Marigold. Ukucubungula ama-amakhulu ezithombe, i-premium ilinganisela kahle.
I-512×512 RGB PNG ene-Spectrum colormap — ekhanyayo / eluhlaza okwesibhakabhaka = eduze, emnyama / eluhlaza okwesibhakabhaka = kude. Amaxabiso e-raw depth ajwayelekile 0-1 phezu kwesithombe. Uma ufuna 16-bit grayscale depth (i-ControlNet noma i-Blender displacement maps), sebenzisa i-API futhi udlulise `format=16bit_png`.
Yebo — i-Marigold output iyifomethi ejwayelekile yokulawula ububanzi-ukungalingani kwe-ControlNet. Sebenzisa ku-Stable Diffusion / FLUX nge-depth ControlNet ukukhombisa kabusha ikhamera eyaziyo: shintsha isiqephu, ukukhangisa, isitayela ngenkathi ugcina isakhiwo se-3D. Free.ai chat ingakuveza wena — sithi "umayelana nobubanzi, bese ubonisa kabusha nge-FLUX ngesitayela se-Studio Ghibli".
I-Marigold iyi-monocular evulekile evulekile ebhekene ne-estimator (CVPR 2024 iphepha elingcono). Kuphakathi kuka-3-5% yeqiniso lendawo ezindaweni zokuhlola zangaphakathi (NYU Depth v2), ngaphakathi kuka-5-8% ngaphandle (KITTI). Ubukhulu be-metric (amamitha e-real-world), hlanganisa nekhamera elawulwa. Ubukhulu obufanayo (lokhu kufutshane kakhulu kunalokho), ukunemba kulungile.
JPG, PNG, WebP kuze kube yi-10MB. I-Marigold isebenza ku-768×768 ngaphakathi kungakhathaliseki ubukhulu be-input futhi iphinde isampule emuva. I-output ihlala i-512×512 RGB. Ubukhulu obuphezulu be-resolution, sebenzisa ukhetho olukhawulelwe olukhawulelwe ngokuqondile.
Yebo — ama-maps abanzi yimingeniso ejwayelekile ye-After Effects 2.5D paralax (i-plugin ye-Camera Mapper), i-DaVinci Resolve esekelwe ekuphazamiseni okubanzi, kanye ne-Blender eshintshanayo. Yenza ububanzi, ungene njenge-alpha-matte sequence, i-keyboard ihamba ngekhamera yakho. Abaphotography abaningi basebenzisa lokhu ukukhangisa i-portrait-mode bokeh kumafoni ngaphandle kwekhamera emibili.
I-Marigold (2024) isebenza kahle kakhulu kune-MiDaS v3 (2022) ne-Intel DPT kuwo wonke ama-benchmark asemthethweni, ikakhulukazi kuma-details angcono (izinwele, i-foliage, ukubukeka okucacile). I-APIs ekhokhelwe (Replicate, fal) ikhokhisa i- $ 0.005-0.02 ngesithombe - sisebenzisa isisindo esifanayo se-Marigold mahhala. Ubukhulu obuhlukile: i-zero.
Yebo — POST ku /v1/image/depth-map/ ngesithombe esiningi. Noma sebenzisa /batch/ ukususa ifolda bese uthola i-ZIP emuva. Kusetshenziswa ukuphatha izinhlelo zefreyimu yevidiyo yonke ukuze kufinyelelwe kumiphumela eyaziwayo yobubanzi.
Yebo — POST ingxenye eminingi `isithombe` ku /v1/isithombe/isithombe-map/ ku api.free.ai. Umsebenzisi ogunyaziwe ngesinqumo sakho se-API. 10,000 amathokheni/inyanga emahhala, ama-levels akhokhelwa phezulu. Ibuyisela i-JSON nge-`image_url` ku-PNG ebanzi. /api/ inesibonelo se-curl.
Thola ku-Photoshop njengendawo yomthombo we-layer-mask, i-ControlNet izici zokubuyisela emuva /image/edit/, i-displacement map ku-Blender, i-paralax animator ku-After Effects, noma i-2.5D dolly-zoom video creator (LeiaPix-style). I-chain ne-/image/edit/ ukudala ububanzi obunezici ezifanele kwithuluzi elinye.
Ungayilinganisela kanjani lethuluzi?