Additional files [RJEE Vol. 9 (1). 2024 – 2]

• Fig. 1. Scheme of data flow and actions for optimal design of AI application in camera trap images processing (semi-automatic classification)
• Fig. 2. Interface of the main components of the Wildlife Insights web-service on the example of the basic module of the CFNR CAMMON project: A – world map of all projects presented on the service; B – main project window with general operational statistics; C – classification window of images grouped into independent events (the number of photos in a given event is shown in the upper left corner of each preview image); D – open registration window with detector and classifier results (possible classification options are shown on the right)
• Fig. 3. Interface of the main components of the Agouti web service on the example of the basic module of the CFNR CAMMON project: A – world map of all projects presented on the platform; B – main window of the project with its description and a map of camera trap locations; C – window of classification of images grouped into independent events (red frame shows available classifier models and classification result); D – window of calibration of camera trap location similar to AnimalTracker web-app
• Fig. 4. Interface of the main components of the WildTrax web service on the example of the basic module of the CFNR CAMMON project: A – world map of all projects presented on the platform; B – organization tab with the list of camera trap locations; C – project tab with the list of species encountered and the number of their images; D – window for classification of images grouped into independent events
• Fig. 5. Timelapse software interface on the example of the basic module of the CFNR CAMMON project. The tagging screen with loaded EcoAssist recognition results is shown. The tagging menu is displayed at the top right; highlighted in green are tags that are filled in automatically when a «quick» tag is clicked or a previous tag is copied. The Quick Paste menu (highlighted with a red frame) allows you to create and select an unlimited number of already prepared «quick» tags
• Fig. 6. Example of animal image detection using calculation of pixel value difference in neighboring images in Timelapse software: A – first frame of roe deer registration “hidden” behind trees; B – second frame of roe deer registration; C – third frame of roe deer registration with triggered detector; D – result of pixel-by-pixel comparison of the third and second frames; E – first frame of European hedgehog registration; F – second frame of hedgehog registration; G – result of pixel-by-pixel comparison of the first two frames with overlay