BG Studio — third-party model and runtime notices ================================================== 1) U2-Net-p segmentation weights (models/u2netp.onnx) - Derived from the U-2-Net project: "U2-Net: Going Deeper with Nested U-Structure for Salient Object Detection" (Qin et al., 2020), https://github.com/xuebinqin/U-2-Net - Licence: Apache License 2.0 (see LICENSE-Apache-2.0.txt in this folder). - Copyright (c) 2020 Xuebin Qin et al. - ONNX export distributed by the rembg project (danielgatis/rembg, MIT) and mirrored on Hugging Face. - Source pinned at: https://huggingface.co/tomjackson2023/rembg/resolve/cd3a3d6767a7859efea31ef0f2f373582cf06d82/u2netp.onnx - SHA-256 (unmodified, byte-identical to the source above): 309c8469258dda742793dce0ebea8e6dd393174f89934733ecc8b14c76f4ddd8 - The weights are shipped UNMODIFIED. They execute entirely inside the visitor's browser; no image data is transmitted anywhere. 1b) MODNet portrait matting weights (models/modnet_fp16.onnx) - Derived from the MODNet project: "MODNet: Real-Time Trimap-Free Portrait Matting via Objective Decomposition" (Ke et al., AAAI 2022), https://github.com/ZHKKKe/MODNet - Licence: Apache License 2.0 (see LICENSE-Apache-2.0.txt in this folder). - Copyright (c) 2020 ZHKKKe (Zhanghan Ke et al.) - fp16 ONNX export by Xenova, mirrored on Hugging Face. Source pinned at: https://huggingface.co/Xenova/modnet/resolve/fa2fa546052fba4c08921230a26cc69a333fca12/onnx/model_fp16.onnx - SHA-256 (unmodified, byte-identical to the source above): 25f165da9bfd30830a575f1f0490f1acd995975cb349bc02f3d79332e1fe5cf6 - The weights are shipped UNMODIFIED and execute entirely inside the visitor's browser; no image data is transmitted anywhere. 2) onnxruntime-web 1.27.0 (ort.wasm.min.js, ort-wasm-simd-threaded.mjs, ort-wasm-simd-threaded.wasm) - https://github.com/microsoft/onnxruntime - Licence: MIT. Copyright (c) Microsoft Corporation. - Shipped unmodified from the onnxruntime-web@1.27.0 npm package. 3) Demo photographs (/demos/*.jpg) - All CC0, from Wikimedia Commons, resized for this app: * couple.jpg — "Happy Couple (Unsplash)" (freestocks.org, CC0) * woman.jpg — "Brunette woman portrait (Unsplash)" (Christopher Campbell, CC0) * hand.jpg — "Setting sun in the West (Unsplash)" (Tim Graf, CC0) * bird.jpg — "Perched blue and orange bird (Unsplash)" (Vincent van Zalinge, CC0) * rose.jpg — "Vibrant red rose on gray (Unsplash)" (Aziz Acharki, CC0) 4) Built-in backdrop photographs (/backdrops/*.jpg) - All CC0, from Wikimedia Commons ("(Unsplash)" collection), resized. Every image was individually reviewed for backdrop suitability (no people, no readable text/brands, no dominant foreground subject): * colorful-1 — "Graffiti on gray wall" (Chris Barbalis) * colorful-2 — "Hot Air Balloon" (Kupono Kuwamura) * colorful-3 — "Colorful hot air balloon in flight" (Aaron Burden) * colorful-4 — "Colorful trees near a lake" (Aaron Burden) * beach-1 — "Beach Waves In The Sand" (Jeremy Cai) * beach-2 — "Palm tree boulevard" (Matthew Hamilton) * beach-3 — "Ice blue waves" (Ahmed Radwan) * sunset-1 — "Open Water During Sunset" (Ruston Youngblood) * sunset-2 — "Golden Hour, Blackpool." (Luke Ellis-Craven) * sunset-3 — "Dramatic golden sunset." (cindy del valle) * sunset-4 — "Colorful sky during sunset" (Rowan Heuvel) * wall-1 — "London brick wall" (Scott Webb) * wall-2 — "Blue Wooden Wall" (Math) * wall-3 — "Old shed wall in close-up" (Zbysiu Rodak) * wall-4 — "Rectangle stone wall" (Reiner Knudsen) * grassland-1 — "Fenced in grass" (Scott Webb) * grassland-2 — "Golden countryside hills" (Adriel Kloppenburg) * grassland-3 — "Rapeseed-blue-sky" (Sweet Ice Cream Photography) * grassland-4 — "Misty hills" (Anthony Ingram) * snow-1 — "Imposing mountain under snow" (Maximilian Wachter) * snow-2 — "Path into the snowy forest" (Artem Kulikov) * snow-3 — "A cabin among the snowy forest" (Gerome Bruneau) * snow-4 — "The Fight between Fall and Winter" (Martin Sattler) * flowers-1 — "Flower trail" (Aaron Brunhofer) * flowers-2 — "Cherry-blossom-japan" (Yuki Yoshida) * flowers-3 — "Beautiful tulip field" (Thomas Cigolla) * street-1 — "City street light trails in Toronto" (Illia Cherednychenko) * street-2 — "City at Night" (Kyle Sudu) * street-3 — "Bright skyline over water" (Osman Rana) * street-4 — "Light trails on a suburban highway" (Joey Kyber) * bokeh-1 — "Bokeh effect" (Gabriel Santiago) * bokeh-2 — "Bokeh the lights of Christmas" (Clem Onojeghuo) * bokeh-3 — "Christmas Lights Bokeh Effect" (thr3 eyes) * bokeh-4 — "Festive Christmas lights create bokeh" (Charlotte Coneybeer) * wood-1 — "Painted square planks" (Chris Barbalis) * wood-2 — "Dark brown firewood texture" (Alex Jones) * wood-3 — "Firewood pile" (Brigitte Tohm) * wood-4 — "Geometric wood pattern" (Teo Duldulao)