In 2025 they are now storing the depth map(s) via concatenating multiple JFIFs after the main image. If you split those you can get the original image and the depth maps from Pixel Camera photos. Source and sample code: https://github.com/dragon66/icafe/issues/90#issuecomment-758031710