In this paper, we present a new motion feature representation for first-person videos. In first-person video understanding (e.g., activity recognition), it is very important to capture both entire scene dynamics (i.e., egomotion) and salient local motion observed in videos. We propose a new feature representation framework based on time series pooling, which is able to abstract detailed short-term/long-term changes in motion/appearance descriptor values over time. The framework is general, handling any type of per-frame descriptor including conventional motion descriptors like histogram of optical flows (HOF) as well as more recent convolutional neural network (CNN) feature descriptors. We experimentally confirm that our approach clearly outperforms previous feature representations including bag-of-visual-words and improved Fisher vector (IFV) when using identical underlying feature descriptors. We also confirm that our feature representation has superior performance to existing state-of-the-art features like local spatio-temporal features and Improved Trajectory Features (originally developed for 3rd-person videos) when handling first-person videos. Multiple first-person activity datasets were tested under various settings to confirm these findings.

本文提出了一种基于时间序列池化的特征表示框架，用于抽象活动视频中的短期/长期特征描述符的变化，并用于处理第一人称视频活动数据集的动作识别问题，与基于视觉词袋和改进的 Fisher 矢量的特征表示相比能够更好地表征第一人称视频中的动作。

第一人称视角视频的动作特征汇聚