A Key Volume Mining Deep Framework for Action Recognition

By Wangjiang Zhu, Jie Hu, Gang Sun, Xudong Cao, Yu Qiao. 2016

240 citations recorded by the source corpus.

Subfields: Artificial Intelligence, Biomedical Engineering, Computer Vision and Pattern Recognition

How researchers see this paper

The researchers whose trusted networks rank it highest, strongest advocate first. Every researcher gets a different order.

In Artificial Intelligence

In Biomedical Engineering

In Computer Vision and Pattern Recognition

Citations in the corpus

RGB-D-based Human Motion Recognition with Deep Learning: A Survey (2017, background): “Based on the two-stream framework, [180] proposed a key volume mining deep framework for action recognition, where they identified key volumes and conducted classification simultaneously.”
Chained Multi-stream Networks Exploiting Pose, Motion, and Appearance for Action Classification and Detection (2017, neutral): “…- Two-stream [33] 88.0% 59.4% - TSN [43] 94.2 % 69.4% - CPD [26] 92.3% 66.2% - Multi-Granular [18] 90.8% 63.6% - M-fusion [28] 89.1% 54.9% - KVMF [49] 93.1% 63.3% - P-CNN [4] - - 61.1% Action tubes [9] - - 62.5% TS R-CNN [29] - - 70.5% MR-TS R-CNN [29] - - 71.1% Ours (chained) 91.1% 69.7% 76.1%…”
Temporal Segment Networks for Action Recognition in Videos (2017, background): “3% KVMF [74] 93 .”
Learning Spatio-Temporal Representation with Pseudo-3D Residual Networks (2017, background): “This architecture is further extended by exploiting multi-granular structure [17, 18, 21], convolutional fusion [6], key-volume mining [39] and temporal segment networks [36] for video representation learning.”
Deep Temporal Linear Encoding Networks (2016, unknown): “…Method UCF101 HMDB51 DT+MVSM [2] 83.5 55.9 iDT+FV [34] 85.9 57.2 Two Stream [24] 88.0 59.4 VideoDarwin [8] − 63.7 C3D [32] 82.3 56.8 Two Stream+LSTM [40] 88.6 − F ST CV (SCI fusion) [29] 88.1 59.1 TDD+FV [36] 90.3 63.2 LTC [33] 91.7 64.8 KVMF [43] 93.1 63.3 TSN [37] 94.0 68.5 3DConv+3DPool [7] 93.”
Temporal Segment Networks: Towards Good Practices for Deep Action Recognition (2016, background): “3% KVMF [41] 93 .”
Hidden Two-Stream Convolutional Networks for Action Recognition (2017, methodology): “We have moved from manually designed features [1], [2], [3], [4], [5] to learned convolutional neural network (CNN) features [6], [7], [8], [9]; from encoding appearance information to encoding motion information [10], [11], [12], [13]; and from learning local features to learning global video featu”
Optical Flow Guided Feature: A Fast and Robust Motion Representation for Video Action Recognition (2017, methodology): “The other compared methods listed in Table 4 include iDT [40], TwoStream ConvNet [29], Two-Stream + LSTM [47], Temporal Deep-convolutional Descriptors (TDD) [41], Long-term Temporal Convolutions (LTC) [37], Key Volume Mining Deep Framework (KVMDF) [52], and also the current stateof-the-art methods s”
A Survey on Deep Learning Based Approaches for Action and Gesture Recognition in Image Sequences (2017, background): “In the same line, [138] introduces a deep-learning method to identify key volumes and classify them simultaneously.”
Dynamic Sampling Networks for Efficient Action Recognition in Videos (2020, background): “[31] proposed KVMF to filter out irrelevant information by identifying key volumes in the sequences.”
Spatiotemporal Multiplier Networks for Video Action Recognition (2017, methodology): “Other recent approaches rely on a Siamese architecture to abstract the temporal transformation of features across a video [39] or identify key volumes in the sequences [43].”
Recognizing Human Actions as the Evolution of Pose Estimation Maps (2018, background): “Recent studies [56, 31, 14, 38, 30] extracted whole human body or body parts instead of whole video for action analysis.”
RPAN: An End-to-End Recurrent Pose-Attention Network for Action Recognition in Videos (2017, methodology): “To improve the performance, several extensions have been proposed by designing trajectory-pooled deep descriptors [45], mining key volume of videos [54], fusing two streams [11], introducing temporal segments [47].”
Spatio-Temporal AutoEncoder for Video Anomaly Detection (2017, background): “As aforementioned, Deep Neural Networks (DNNs) have been proven superior in discriminative computer vision tasks, such as image classi cation [9], object detection [21], semantic segmentation [10], as well as a serious of video analysis problems [26, 30, 31, 35].”
ActionVLAD: Learning Spatio-Temporal Aggregation for Action Classification (2017, background): “CNN-based representations [15, 51, 58, 59, 63] have not yet significantly outperformed the best hand-engineered descriptors [12, 53].”