KiSSHoer

When infants are born with hearing impairments, early detection and care is essential. It also plays an important role in the later acquisition of hearing and spoken language. Current methods of speech diagnostics are not sufficiently oriented towards this target group and are very time-consuming, as they require phonetic transcriptions and phoneme-specific error analysis. For this reason, researchers at the Fraunhofer IDMT in Oldenburg are working on a time-saving evaluation aid in the project “Children's speech screening for hearing disorders” (KiSSHoer). The aim is to develop an automated pronunciation assessment for children between the ages of three and five.

The use of speech processing technologies should make it possible to carry out the screening automatically and thus reduce the time required by experts in the diagnostic process. In addition to the derivation of individual therapy goals, progress and changes in the level of development should be objectively and efficiently presented. This applies to both speech therapy and technical care, for example through adjustments to hearing aids or cochlear implants.

In terms of technology, KiSSHoer is based on analysis functions and technologies from the “Oldenburg Speech and Voice Test” (OSST) and “Speech and Voice Analysis Technologies as a Service” (SVAT) projects.