Multi-scale Deep Learning for Gesture Detection and Localization

By Natalia Neverova, Christian Wolf, Graham W. Taylor, Florian Nebout. 2014

233 citations recorded by the source corpus.

Subfields: Biomedical Engineering, Computer Vision and Pattern Recognition, Human-Computer Interaction

How researchers see this paper

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

In Biomedical Engineering

1. Changshui Zhang
2. Hu Liu
3. Runpeng Cui
4. Sergio Escalera
5. Shuai Zhou
6. S. Li
7. Xiaoyuan Guo
8. Jiali Duan

In Computer Vision and Pattern Recognition

1. Dacheng Tao
2. Luc Van Gool
3. Xiaogang Wang
4. Yi Yang
5. Trevor Darrell
6. Chen Change Loy
7. Ming-Hsuan Yang
8. Wanli Ouyang

In Human-Computer Interaction

1. Sergio Escalera
2. Isabelle M Guyon
3. Changshui Zhang
4. Hu Liu
5. Runpeng Cui
6. Jun Wan
7. Shuai Zhou
8. Peiyi Shen

Citations in the corpus

• Convolutional Neural Network for Paraphrase Identification (2015, background): “Inspired by recent successes of deep neural networks (NNs) in fields like computer vision (Neverova et al., 2014), speech recognition (Deng et al., 2013) and natural language processing (Col-lobert and Weston, 2008), we adopt a deep learning approach to paraphrase identification in this paper.”
• Recurrent Convolutional Neural Networks for Continuous Sign Language Recognition by Staged Optimization (2017, methodology): “The neural network methods adopted in gesture analysis include CNNs [17, 18, 26], 3D CNNs [21, 23, 30] and temporal convolutions [26].”
• Interacting with Soli: Exploring Fine-Grained Dynamic Gesture Recognition in the Radio-Frequency Spectrum (2016, background): “Similarly [24] uses CNNs for sign language recognition based on combined color and depth data.”
• Feature Fusion for Online Mutual Knowledge Distillation (2019, methodology): “From this fact, researchers found that combining the features of each layer increases the performance of the model and showed the effectiveness of this method in various computer vision tasks such as detection, semantic segmentation and gesture classification [22, 10, 7, 4].”
• Hand gesture recognition with 3D convolutional neural networks (2015, methodology): “However, their technique was intended for gestures performed indoors only.”
• Large-scale Isolated Gesture Recognition using pyramidal 3D convolutional networks (2016, methodology): “The multi-scale deep learning method [3] takes samples with three different durations to fulfill the multi-scale deep neural networks.”
• A Unified Framework for Multi-Modal Isolated Gesture Recognition (2018, background): “Neverova et al [Neverova et al. 2015] proposes a multi-modal architecture that operates at 3 temporal scales corresponding to dynamic poses for gesture localization.”
• ChaLearn looking at people: A review of events and resources (2017, background): “7173 (Accuracy) [51] Gesture Recognition ChaLearn LAP Workshop ECCV 2014 Multimodal Gesture Recognition: Montalbano V2 JMLR Gesture 2014, TPAMI HuPBA 2016 and IJCV LAP 2016 LIRIS: 0.849987 (Accuracy) [52] Action/Interaction Recognition (Second round) ChaLearn LAP Workshop CVPR 2015 Action/Interactio”
• Multi-Modality Fusion based on Consensus-Voting and 3D Convolution for Isolated Gesture Recognition (2016, background): “Neverova et al [16] proposes a multi-modal architecture that operates at 3 temporal scales corresponding to dynamic poses for gesture localization.”
• Learning Multiscale Features Directly from Waveforms (2016, methodology): “Multiscale convolutions have already been successfully applied to address tasks in the computer vision field, such as image classification [9], scene labeling [10], and gesture detection [11].”
• A Low Power, Fully Event-Based Gesture Recognition System (2017, background): “Challenging lighting conditions constitute another confounding variable, and are typically addressed by fusing data from different sensing modalities [32], increasing power consumption.”
• Towards Deep Developmental Learning (2016, background): “Here again, a lot of effort is dedicated to these concerns in the deep learning literature, with the same methodological limitations as above [85], [86].”
• A Survey on Deep Learning Based Approaches for Action and Gesture Recognition in Image Sequences (2017, methodology): “Advances in hybrid models combining handcrafted and new descriptors are expected [68, 112].”
• Challenges in multimodal gesture recognition (2016, methodology): “The winner of the competition was also the work that won the ChaLearn challenge at ECCV 2014 (Neverova et al., 2014b).”
• Whole-Body Human Pose Estimation in the Wild (2020, methodology): “For hand keypoint estimation, most work rely on auxiliary information such as depth information [40,50,52] or multi-view [15,36] information.”