Annotated Web Ears Toolbox – AWE Toolbox

The toolbox is written in Matlab. You need Mex for it to run. The toolbox already includes Annotated Web Ears Database. It should work out of the box if you have environment set correctly. If it does not work, see the help section.

Annotated Web Ears Dataset for segmentation – AWE Dataset4Seg

  • Download will be available after the paper is published.
The database contains 1.000 images of 100 persons with the pixel-wise annotations of ear locations.

Annotated Web Ears Dataset – AWE Dataset

The database contains 1.000 images of 100 persons. Images were taken from the wild (acquired from the internet) and contain the following annotations: gender, ethnicity, accessories, occlusions, head pitch, head roll, head yaw, head side, central tragus point – all stored in JSON files.

The Additional Dataset

This database contains 804 images of 16 persons. Images are taken from the wild (acquired from the internet). Persons and images are different to AWE dataset, and annotations are not as detailed: ear direction and central tragus points stored in JSON files. You can use this dataset for some preliminary testing or for help during the development.

Annotated Web Ears Toolbox Only

Only the toolbox, without the datasets.

Ear Cropper & Annotation Tool

Do you want to create your own ear dataset? Use this tool to crop ears out of images and add annotations in a quick way. You need a working server running PHP in order to run the tool.