ClawCraneNet: Leveraging Object-level Relation for Text-based Video Segmentation
This addresses the challenge of segmenting objects in videos based on natural language descriptions, which is crucial for applications like video editing and robotics, though it is an incremental improvement over existing methods.
The paper tackles text-based video segmentation by proposing a top-down approach that mimics human reasoning, using object-level relations to select referred objects, and achieves state-of-the-art performance on A2D Sentences and J-HMDB Sentences datasets.
Text-based video segmentation is a challenging task that segments out the natural language referred objects in videos. It essentially requires semantic comprehension and fine-grained video understanding. Existing methods introduce language representation into segmentation models in a bottom-up manner, which merely conducts vision-language interaction within local receptive fields of ConvNets. We argue that such interaction is not fulfilled since the model can barely construct region-level relationships given partial observations, which is contrary to the description logic of natural language/referring expressions. In fact, people usually describe a target object using relations with other objects, which may not be easily understood without seeing the whole video. To address the issue, we introduce a novel top-down approach by imitating how we human segment an object with the language guidance. We first figure out all candidate objects in videos and then choose the refereed one by parsing relations among those high-level objects. Three kinds of object-level relations are investigated for precise relationship understanding, i.e., positional relation, text-guided semantic relation, and temporal relation. Extensive experiments on A2D Sentences and J-HMDB Sentences show our method outperforms state-of-the-art methods by a large margin. Qualitative results also show our results are more explainable.