Can Multimodal LLMs do Visual Temporal Understanding and Reasoning? The answer is No!
This work addresses a critical gap in MLLM capabilities for real-world dynamic comprehension, but it is incremental as it primarily introduces a new benchmark without proposing novel solutions.
The paper tackles the problem of evaluating multimodal large language models (MLLMs) on visual temporal understanding and reasoning, revealing that advanced models like GPT-4o achieve only 49.1% accuracy in temporal order tasks and 70% in time-lapse estimation, indicating significant limitations.
Multimodal Large Language Models (MLLMs) have achieved significant advancements in tasks like Visual Question Answering (VQA) by leveraging foundational Large Language Models (LLMs). However, their abilities in specific areas such as visual temporal understanding, which is crucial for comprehending real-world dynamics, remain underexplored. To address this, we propose a challenging evaluation benchmark named TemporalVQA, consisting of two parts: 1) Temporal Order Understanding and 2) Time-lapse Estimation. The first part requires MLLMs to determine the sequence of events by analyzing temporally consecutive video frames. The second part presents image pairs with varying time differences, framed as multiple-choice questions, asking MLLMs to estimate the time-lapse between images with options ranging from seconds to years. Our evaluations of advanced MLLMs, including models like GPT-4o and Gemini-1.5-Pro, reveal significant challenges: GPT-4o achieved only 49.1% average consistent accuracy in temporal order task and 70% in time-lapse estimation, with open-source models performing even poorly. These findings underscore the limitations of current MLLMs in visual temporal understanding and reasoning, highlighting the need for further improvements for their temporal capability. Our dataset can be found at https://huggingface.co/datasets/fazliimam/temporal-vqa.