Multimodal Event Detection: Current Approaches and Defining the New Playground through LLMs and VLMs