Catch Your Emotion: Sharpening Emotion Perception in Multimodal Large Language Models

Multimodal large language models (MLLMs) have achieved impressive progress in tasks such as visual question answering and visual understanding, but they still face significant challenges in emotional reasoning. Current methods to enhance emotional understanding typically rely on finetuning or manual annotations, which are resourceintensive and limit scalability. In this work, we focus on improving the ability of MLLMs to capture emotions during the inference phase. Specifically, MLLMs encounter two main issues in the inference stage: they struggle to distinguish between semantically similar emotions, leading to misclassification, and they are overwhelmed by redundant or irrelevant visual information, which distracts from key emotional cues. To address these, we propose a training-free method named Sharpening Emotion Perception in MLLMs (SEPM), which incorporates a Confidence-Guided Coarseto-Fine Inference framework to refine emotion classification by guiding the model through simpler tasks. Additionally, SEPM employs Focuson-Emotion Visual Augmentation to reduce visual redundancy by directing the attention of models to relevant emotional cues in images. Experimental results demonstrate that SEPM significantly improves MLLM performance on emotionrelated tasks, providing a resource-efficient and scalable solution for emotion recognition. Our code is available in https://github.com/ fuyyyyy/SEPM .