AI Paper Atlas

Research map / Vision-language and multimodal / Multimodal LLMs

EMO-LLaMA: Enhancing Facial Emotion Understanding with Instruction Tuning

Bohao Xing, Zitong Yu, Xin Liu, Kaishen Yuan, Qilang Ye, Weicheng Xie, Huanjing Yue, Jingyu Yang, Heikki A. Kalviainen

IJCV 2026 · Vision-language and multimodal › Multimodal LLMs · approach: Benchmarks and analysis of MLLMs · 50 citations (OpenAlex)

Read the paperCheck your idea against 137,217 papers

AI-extracted. The measured effects below were read by AI from the paper's full text (no summary yet). Numbers and claims can be wrong or missing context, so check the paper before you cite or build on it.

Measured effects reported in the paper

What was changedWhereMeasuredEffect
add facial embedding and ROI landmark priors (ablation)EMO-LLaMA · MAFWUAR: 34.75 -> 40.67Improved
add full landmark embedding (ablation)EMO-LLaMA · MAFWUAR: 40.67 -> 41.14Improved
add age-gender-race handcrafted prompt (ablation)EMO-LLaMA · MAFWUAR: 41.14 -> 41.57Improved
use Face Info Mining instead of pool and concat (ablation)EMO-LLaMA · MAFWUAR: 34.51 -> 40.67Improved
add category instruction data (ablation)LLaMA-VID · AffectNetaccuracy: 32.96 -> 57.93Improved
add conversation instruction data (ablation)EMO-LLaMA · AffectNetaccuracy: 57.93 -> 59.58Improved
add conversation instruction data (ablation)EMO-LLaMA · MAFWUAR: 31.56 -> 34.75Improved
add descriptive text at inference (ablation)EMO-LLaMA · MAFWUAR: 41.57 -> 49.29Improved
add imbalanced dataset sampling (ablation)EMO-LLaMA · AffectNetaccuracy: 11.30 -> 57.93Improved
add imbalanced dataset sampling (ablation)EMO-LLaMA · MAFWWAR: 39.27 -> 44.15Improved
train on image data and test zero-shot on video versus GPT-4VEMO-LLaMA · DFEWUAR: 36.96 -> 44.73Improved
train on video data and test zero-shot on images versus GPT-4VEMO-LLaMA · AffectNetaccuracy: 42.77 -> 38.43Hurt

Closest papers by meaning

Writing a paper in this area? scime draft is a free LaTeX editor with AI included, real-time co-authors and a check of your references and venue rules before you submit. It also imports OpenAI Prism projects.

Open scime draft

Get the next scime Dispatch and a monthly field update

Free. We send a confirmation link first; every email has a one-click unsubscribe. Privacy