A spectrogram and local feature-assisted convolutional neural network for Amharic speech emotion identification

Yeshambel Asmare Mengist, Abrham Debasu Mengistu, Mulatu Yirga Beyene, Mikiyas Assefa Kassa, Getasew Asmare Mengist · 2026

Read A spectrogram and local feature-assisted convolutional neural network for Amharic speech emotion identification by Yeshambel Asmare Mengist, Abrham Debasu Mengistu, Mulatu Yirga Beyene, Mikiyas Assefa Kassa, Getasew Asmare Mengist (2026) in the Tower of Babel, a free online library of public-domain works preserved in clean, searchable text. No sign-up. The full text is served as a plain-text file at a predictable URL, where text is present: https://antarctic-labs.com/tower-of-babel/pmc-13346804.txt.

Description

Yeshambel Asmare Mengist and co-authors built in 2026 a speech emotion spotter for Amharic, a language with few data sets where emotion spotting lags. Better emotion spotting could lift both human-computer links and talk aids. Yet most systems train on well-resourced languages.

The team made a data set of 1,650 three-second Amharic clips labelled as anger, fear, joy, neutral, or sad. They joined spectrogram images with local sound traits such as pitch, energy, and cepstral coefficients in a neural net. Spectral cleaning and wavelet denoising sharpened the signals. The model shows how careful trait design can bring emotion-aware tools to languages that mainstream work often skips.

Catalog record

CREATORYeshambel Asmare Mengist, Abrham Debasu Mengistu, Mulatu Yirga Beyene, Mikiyas Assefa Kassa, Getasew Asmare Mengist
YEAR2026
COLLECTIONPAPERS
FORMATTXT
FILE SIZE40 KB
VERSION1.0
LICENSECC BY 4.0
RIGHTSCC BY
SOURCEEurope PMC

Source: https://europepmc.org/articles/PMC13346804

Browse the full library · About the Tower of Babel