SeriesBench: A Benchmark for Narrative-Driven Drama Series Understanding
This addresses the need for better evaluation of MLLMs on complex, continuous narratives in video series, which is incremental as it builds on existing video understanding benchmarks by adding narrative-specific tasks.
The authors tackled the problem of evaluating Multi-modal Large Language Models (MLLMs) on narrative-driven drama series, which existing benchmarks overlook by focusing on standalone videos, and proposed SeriesBench, a benchmark with 105 series and 28 tasks, along with PC-DCoT, a narrative reasoning framework that improves MLLM performance on these tasks.
With the rapid development of Multi-modal Large Language Models (MLLMs), an increasing number of benchmarks have been established to evaluate the video understanding capabilities of these models. However, these benchmarks focus on standalone videos and mainly assess "visual elements" like human actions and object states. In reality, contemporary videos often encompass complex and continuous narratives, typically presented as a series. To address this challenge, we propose SeriesBench, a benchmark consisting of 105 carefully curated narrative-driven series, covering 28 specialized tasks that require deep narrative understanding. Specifically, we first select a diverse set of drama series spanning various genres. Then, we introduce a novel long-span narrative annotation method, combined with a full-information transformation approach to convert manual annotations into diverse task formats. To further enhance model capacity for detailed analysis of plot structures and character relationships within series, we propose a novel narrative reasoning framework, PC-DCoT. Extensive results on SeriesBench indicate that existing MLLMs still face significant challenges in understanding narrative-driven series, while PC-DCoT enables these MLLMs to achieve performance improvements. Overall, our SeriesBench and PC-DCoT highlight the critical necessity of advancing model capabilities to understand narrative-driven series, guiding the future development of MLLMs. SeriesBench is publicly available at https://github.com/zackhxn/SeriesBench-CVPR2025.