Automatic emotion recognition is becoming increasingly vital across various sectors where emotion analysis plays a crucial role. This technology has applications in customer service, marketing, healthcare, education, automotive, entertainment, and security, allowing for real-time emotional insights that enhance user experiences and optimize system responses. Despite its growing significance, emotion recognition remains challenging due to the complexity of emotions, cultural variability, and the technological barriers associated with accurate classification. This paper provides a scoping review of state-of-the-art methods in machine emotion recognition, covering capabilities such as facial expression-based emotion recognition (FER), eye-tracking, pupillometry, electrooculography (EOG), microexpression analysis, and gait analysis. The review also explores body language and posture recognition, hand gestures, touch dynamics, wearable devices, and speech and text analysis, including spoken and written text. Furthermore, physiological signal-based methods, including respiration rate analysis (RR), galvanic skin response (GSR), electroencephalography (EEG), electromyography (EMG), skin temperature measurements (SKT), electrocardiography (ECG/EKG), and heart rate variability (HRV), are discussed. The paper also highlights the challenges encountered, including data limitations, computational complexity, and the subjectivity and context-dependency of emotions. This comprehensive review can provide a foundation for further exploration into artificial intelligence-based emotion recognition.