The field of emotion recognition and multimodal interaction is moving towards more inclusive and diverse solutions, with a focus on developing intelligent educational tools and accessible technologies for individuals with disabilities. Recent studies have explored the use of sign language recognition, affect mining techniques, and multimodal datasets to improve emotion understanding and recognition. Noteworthy papers in this area include the development of a Palestinian sign language recognition system, which achieved an accuracy of 97.59% in recognizing mathematical signs, and the introduction of EmoSign, a multimodal dataset for understanding emotions in American Sign Language. The EmotionTalk dataset, which provides rich annotations for Chinese multimodal emotion recognition, is also a significant contribution to the field. Additionally, research on head motion patterns as generalisable depression biomarkers and the development of supervised contrastive learning approaches, such as EmotionRankCLAP, demonstrate the innovative and advancing nature of this field.