The original `dog.jpg` is **Golde33443.jpg**, by **Golden Trvs Gol twister**, dated 26 February 2012, licensed **CC BY-SA 3.0 Unported**. - Source: https://commons.wikimedia.org/wiki/File:Golde33443.jpg - Original: https://upload.wikimedia.org/wikipedia/commons/6/6e/Golde33443.jpg - License: https://creativecommons.org/licenses/by-sa/3.0/ The original photograph is unmodified. `tools/prepare-tests.ps1` creates an adaptation: a padded 640×360, 12-second still-video with a **synthetic test tone**. That generated video is also CC BY-SA 3.0, is ignored by Git, and is a labeled test/demo fixture. It is never substituted for a configured camera. The model must infer the dog from pixels; tests do not inject its detection. The line-crossing integration test also makes a temporary animated adaptation of this photograph moving across a black frame. That adaptation is CC BY-SA 3.0 as well. Its detections and tracking are generated by the real model; known detection objects are not injected into that integration test. Additional test-only photos are declared with original URLs and SHA-256 hashes in `classes.json`. Each source page identifies its author's CC0 1.0 dedication: - Cat: [Picture of cat.jpg](https://commons.wikimedia.org/wiki/File:Picture_of_cat.jpg), Theeyes 07. - Person: [A man doing Isometric walking.jpg](https://commons.wikimedia.org/wiki/File:A_man_doing_Isometric_walking.jpg), Vlad Kozh. - Car: [Toyota Mirai rally car front](https://commons.wikimedia.org/wiki/File:Toyota_Mirai_rally_car_front_-_Tokyo_Auto_Salon_2015.jpg), Loadbeta. License: https://creativecommons.org/publicdomain/zero/1.0/. These photos and their generated 12-second padded still-video adaptations stay in `.tools` and are not included in application installers. `tools/test-detection-classes.ps1` uses the actual published worker, its checksum-verified model, camera streams, rules, snapshots and finalized recordings; it never injects detections. Passing these known fixtures demonstrates the exercised class pipelines, not broad field accuracy or identification of any depicted person. WEBSITE SCREENSHOT ADAPTATIONS The attribution above is copied from the Vigivo application test fixture notes. On this website the photographs are visible within native app screenshots. Their presentation may include fitting, cropping, detection boxes, labels and interface overlays. Photo adaptations of Golde33443.jpg retain CC BY-SA 3.0. The original Vigivo interface, product branding and logos remain proprietary to Asia Design Lab. These screenshots do not contain customer camera footage. TYPOGRAPHY Copyright 2018 The Manrope Project Authors https://github.com/googlefonts/manrope The locally hosted Manrope font is licensed under the SIL Open Font License 1.1. The complete license is supplied alongside this file as fonts-OFL.txt. PRODUCT BRANDING Vigivo AI, ADL Messenger and Animail logos and product branding: Asia Design Lab. https://www.asiadesignlab.com/