Efﬁcient feature extraction, encoding and classiﬁcation for action recognition

By Vadim Kantorov, Ivan Laptev

Citations in the corpus

Explainable Artificial Intelligence: Understanding, Visualizing and Interpreting Deep Learning Models (2017, neutral): “The last examples demonstrates the explanation of a Fisher Vector / SVM classiﬁer [16], which was trained for predicting human actions from compressed videos.”
Long-Term Temporal Convolutions for Action Recognition (2016, neutral): “To investigate the dependency of action recognition on the quality of motion estimation, we experiment with three types of ﬂow inputs obtained either directly from the video encoding, referred to as MPEG ﬂow [27], or from two optical ﬂow estimators, namely Farneback [28] and Brox [29].”
End-to-End Learning of Action Detection from Frame Glimpses in Videos (2015, neutral): “There is a long history of work in video analysis and activity recognition [20, 49, 2, 31, 17, 8, 10, 12, 50].”
ECO: Efficient Convolutional Network for Online Video Understanding (2018, unknown): “Kantorov [26] introduced a fast dense trajectory method.”
An overview of deep learning based methods for unsupervised and semi-supervised anomaly detection in videos (2018, unknown): “• Action recognition is an important domain in computer vision which requires supervised learning of efﬁcient representations of appearance and motion for the purpose of classiﬁcation [21].”
Hidden Two-Stream Convolutional Networks for Action Recognition (2017, unknown): “This indicates that our stacked end-to-end learning framework can implicitly learn better motion representations than motion vectors [10, 33] and RGB diﬀerences [25] with respect to the task of action recognition.”
MARS: Motion-Augmented RGB Stream for Action Recognition (2019, unknown): “We also evaluate the performance of two computationally efﬁcient ﬂow approaches, MPEGFlow [17] and PWC-Net [34] on MiniKinetics.”
Hide-and-Seek: Forcing a Network to be Meticulous for Weakly-Supervised Object and Action Localization (2017, unknown): “Action localization is a well studied problem [28, 6, 50, 20, 22].”