Question-guided Visual Compression with Memory Feedback for Long-Term Video Understanding

In the context of long-term video understanding with large multimodal models, many frameworks have been proposed. Although transformer-based visual compressors and memory-augmented approaches are often used to process long videos, they usually compress each frame independently and therefore fail to achieve strong performance on tasks that require understanding complete events, such as temporal ordering tasks in MLVU and VNBench. This motivates us to rethink the conventional one-way scheme from perception to memory, and instead establish a feedbackdriven process in which past visual contexts stored in the context memory can benefit ongoing perception. To this end, we propose Question-guided Visual Compression with Memory Feedback (QViC-MF), a framework for long-term video understanding. At its core is a Question-guided Multimodal Selective Attention (QMSA), which learns to preserve visual information related to the given question from both the current clip and the past related frames from the memory. The compressor and memory feedback work iteratively for each clip of the entire video. This simple yet effective design yields large performance gains on longterm video understanding tasks. Extensive experiments show that our method achieves significant improvement over current state-of-the-art methods by 6.1% on MLVU test, 8.3% on LVBench, 18.3% on VNBench Long, and 3.7% on VideoMME Long. The code will be released publicly.

IQViC: In-context, Question Adaptive Vision Compressor for Long-term Video Understanding LMMs

With the increasing complexity of video data and the need for more efficient long-term temporal understanding, existing long-term video understanding methods often fail to accurately capture and analyze extended video sequences. These methods typically struggle to maintain performance over longer durations and to handle the intricate dependencies within the video content. To address these limitations, we propose a simple yet effective large multi-modal model framework for long-term video understanding that incorporates a novel visual compressor, the In-context, Question Adaptive Visual Compressor (IQViC). The key idea, inspired by humans' selective attention and in-context memory mechanisms, is to introduce a novel visual compressor and incorporate efficient memory management techniques to enhance long-term video question answering. Our framework utilizes IQViC, a transformer-based visual compressor, enabling question-conditioned in-context compression, unlike existing methods that rely on full video visual features. This selectively extracts relevant information, significantly reducing memory token requirements. Through extensive experiments on a new dataset based on InfiniBench for long-term video understanding, and standard benchmarks used for existing methods' evaluation, we demonstrate the effectiveness of our proposed IQViC framework and its superiority over state-of-the-art methods in terms of video understanding accuracy and memory efficiency.

A Unified Multi-view Multi-person Tracking Framework

Although there is a significant development in 3D Multi-view Multi-person Tracking (3D MM-Tracking), current 3D MM-Tracking frameworks are designed separately for footprint and pose tracking. Specifically, frameworks designed for footprint tracking cannot be utilized in 3D pose tracking, because they directly obtain 3D positions on the ground plane with a homography projection, which is inapplicable to 3D poses above the ground. In contrast, frameworks designed for pose tracking generally isolate multi-view and multi-frame associations and may not be robust to footprint tracking, since footprint tracking utilizes fewer key points than pose tracking, which weakens multi-view association cues in a single frame. This study presents a Unified Multi-view Multi-person Tracking framework to bridge the gap between footprint tracking and pose tracking. Without additional modifications, the framework can adopt monocular 2D bounding boxes and 2D poses as the input to produce robust 3D trajectories for multiple persons. Importantly, multi-frame and multi-view information are jointly employed to improve the performance of association and triangulation. The effectiveness of our framework is verified by accomplishing state-of-the-art performance on the Campus and Shelf datasets for 3D pose tracking, and by comparable results on the WILDTRACK and MMPTRACK datasets for 3D footprint tracking.