Target Images


The figure you are augmenting we call a 'target image,' because mobile devices find and track the 'target' through their camera.

Some images make much better targets than others. After you upload your target image, it will be scored as to how good it is at being found and tracked.

If your figure receives a low score, your augmentations may appear shakey or not show up at all.

Tips for a Good Target Image

  • Contrast is the #1 most important factor. Try increasing the contrast of your image.

  • Color dosn't matter. Only contrast does.

  • Images that are repetitive or 'sparse' are difficult to detect and track.

  • Increasing the resolution of images does not improve tracking, but you don't want to be below 300 x 300 pixels.

  • A subset of your figure may track better.

  • Including some additional details around your figure such as the figure caption or additional boundary lines can, in some cases, improve tracking.

Things to Keep in Mind

  • Use the exact image you upload to Schol-AR in your publication or project. It is tracked on a pixel-by-pixel level, and any changes will negatively impact detection and tracking.

  • Printing is not required for Schol-AR to function. Images are detected and tracked perfectly well on digital screens.

  • If the target image is printed, and becomes wrinkled or bent, it will track poorly.

  • In general mobile devices will only detect these images when they fill ~25% of their cameras screen. So you cannot detect images that are far away.



Updates

Contact: Schol-AR@ardist.io