Utilivo Photo: third-party components

ONNX Runtime Web 1.31.1
Copyright (c) Microsoft Corporation. MIT License, see ort/LICENSE.txt.
Original distribution: https://registry.npmjs.org/onnxruntime-web/-/onnxruntime-web-1.31.1.tgz
Selected JavaScript and WASM files are redistributed unchanged.
Upstream: https://github.com/microsoft/onnxruntime

U²-NetP model (ONNX)
Authors: Xuebin Qin, Zichen Zhang, Chenyang Huang, Masood Dehghan,
Osmar R. Zaiane, Martin Jagersand.
Paper: U²-Net: Going Deeper with Nested U-Structure for Salient Object Detection.
Upstream: https://github.com/xuebinqin/U-2-Net
License: Apache License 2.0, see u2net/LICENSE.txt.
Author's model-license clarification: https://github.com/xuebinqin/U-2-Net/issues/208
ONNX conversion distributed by rembg: https://github.com/danielgatis/rembg
Model bytes copied without modification from https://huggingface.co/edgetools/u2netp
and checked against the checksum pinned by rembg:
MD5: 8e83ca70e441ab06c318d82300c84806
SHA-256: 309c8469258dda742793dce0ebea8e6dd393174f89934733ecc8b14c76f4ddd8

Preprocessing and mask normalization were independently implemented following
the public rembg U2netpSession description. No rembg source is bundled.
This is a general object-segmentation model, not an identity or ICAO validator.
No biometric identification or facial retouching is performed by this model.