Rethinking Video Segmentation with Masked Video Consistency: Did the Model Learn as Intended?
This addresses inconsistent segmentation for video analysis applications, representing a strong incremental improvement.
The paper tackled inconsistent video segmentation from small or imbalanced datasets by proposing Masked Video Consistency and Object Masked Attention, achieving state-of-the-art performance on five datasets for three tasks without adding parameters.
Video segmentation aims at partitioning video sequences into meaningful segments based on objects or regions of interest within frames. Current video segmentation models are often derived from image segmentation techniques, which struggle to cope with small-scale or class-imbalanced video datasets. This leads to inconsistent segmentation results across frames. To address these issues, we propose a training strategy Masked Video Consistency, which enhances spatial and temporal feature aggregation. MVC introduces a training strategy that randomly masks image patches, compelling the network to predict the entire semantic segmentation, thus improving contextual information integration. Additionally, we introduce Object Masked Attention (OMA) to optimize the cross-attention mechanism by reducing the impact of irrelevant queries, thereby enhancing temporal modeling capabilities. Our approach, integrated into the latest decoupled universal video segmentation framework, achieves state-of-the-art performance across five datasets for three video segmentation tasks, demonstrating significant improvements over previous methods without increasing model parameters.