KIDO Dataset is Now Publicly Available: A Multimodal Dataset for Children's Emotion Analysis

[Serdar ÇİFTÇİ]

Dear Colleagues,

I am pleased to share that our latest dataset, KIDO (Kids' Drawings for Emotion Analysis), is now publicly available to the research community.

KIDO is a multimodal dataset consisting of 5,430 children's hand-drawn drawings together with their self-reflection texts, in which children explain why their drawing represents happiness or sadness. In addition to the drawings and textual reflections, the dataset also includes emotion labels and demographic information, making it a valuable resource for multimodal learning.

To the best of our knowledge, KIDO is one of the first publicly available datasets that jointly combines children's drawings with their accompanying written reflections. We believe it can facilitate research across several fields, including computer vision, multimodal learning, affective computing, psychology, educational technologies, natural language processing, and human-centered AI.

Our paper:
https://ieeexplore.ieee.org/abstract/document/11151239

GitHub repository:
https://github.com/serdarciftci/KIDO

We hope KIDO will be useful for the community and inspire further research in multimodal emotion understanding. 

Thank you for your interest, and we welcome any feedback or collaboration.

Best regards,

-- 
Serdar Ciftci, Ph.D.
Assistant Professor
Harran University
Department of Computer Engineering
https://user.ceng.metu.edu.tr/~sciftci