Memotion Analysis: Multimodal Fusion Techniques for Humor Classification in Memes

DOI

10.2991/978-94-6239-664-7_73How to use a DOI?

Keywords

Multimodal Learning; Meme Classification; Fea ture Concatenation; BiLSTM; Deep Learning; Transfer Learning

Abstract

This paper presents a multimodal deep learning framework for humor classification in memes, leveraging both textual and visual information to improve sentiment understanding in internet content. The study explores unimodal and multimodal configurations by integrating image-based CNN architectures (MobileNetV2, ResNet152, YOLOv4, VGG19) with text based models (BiLSTM) through feature fusion techniques. To address class imbalance and overfitting, the dataset—comprising 6,982 labeled memes—was balanced via augmentation and sample equalization across four humor levels: Not Funny, Funny, Very Funny, and Hilarious. Among the various fusion strategies, the MobileNetV2–BiLSTM combination achieved the highest performance, with a precision of 89% and an F1-score of 80% on the test subset. However, performance declined on the full dataset due to increased variance and minority class underrepresentation. The findings highlight the importance of modality interaction, feature fusion, and dataset balancing in achieving robust and interpretable meme sentiment analysis. Future directions include adaptive fusion mechanisms, transformer based architectures, and domain-specific knowledge integration to enhance generalization and contextual sensitivity in humor classification.

Copyright

© 2026 The Author(s)

Open Access

Open Access This chapter is licensed under the terms of the Creative Commons Attribution-NonCommercial 4.0 International License (http://creativecommons.org/licenses/by-nc/4.0/), which permits any noncommercial use, sharing, adaptation, distribution and reproduction in any medium or format, as long as you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons license and indicate if changes were made.

Download article (PDF)

TY  - CONF
AU  - Aisha Tasnim Aishy
AU  - Samia Halim Zanvi
AU  - Md Sowkat Ali
AU  - Mohammed Maruf Hossen
AU  - M. Shahriar Mahmud Rafi
AU  - Md Ashraful Islam
PY  - 2026
DA  - 2026/06/08
TI  - Memotion Analysis: Multimodal Fusion Techniques for Humor Classification in Memes
BT  - Proceedings of the International Conference on Intelligent Data Analysis and Applications (IDAA 2025)
PB  - Atlantis Press
SP  - 1066
EP  - 1078
SN  - 1951-6851
UR  - https://doi.org/10.2991/978-94-6239-664-7_73
DO  - 10.2991/978-94-6239-664-7_73
ID  - Aishy2026
ER  -