2019 · 36 citations · 25 references
ComputingComputer ScienceInjury PreventionComputer VisionMachine LearningMovement AnalysisAction RecognitionDeep LearningFall DetectionRepresentation LearningTransfer LearningTransfer TrainingAccelerometryContinuous Wavelet TransformConvolutional Neural NetworksData AugmentationDeep Learning For Fall DetectionMachine Learning For Fall DetectionMusculoskeletal ModelingTransfer Learning For Fall Detection
This paper presents early work on a fall detection method using transfer learning method, in conjunction with a long-term effort to combine efficient machine learning and prior personalized musculoskeletal modeling to deploy fall injury mitigation in geriatric subjects. Inspired by the tremendous progress in image-based object recognition with deep convolutional neural networks (DCNNs), we opt for a pre-trained kinematics-based machine learning approach through existing large-scale annotated accelerometry datasets. The accelerometry datasets are converted to images using time-frequency analysis, based on scalograms, by computing the continuous wavelet transform filter bank. Subsequently, data augmentation is performed on these scalogram images to increase accuracy, thereby complementing limited labeled fall sensor data, enabling transfer learning from the existing pre-trained model. The experimental results on publicly available URFD datasets demonstrate that transfer learning leads to a better performance than the existing methods in the case of scarce labeled training data.
25
ImageNet: A large-scale hierarchical image database
Jia Deng, Wei Dong, Richard Socher et al. · 2009 IEEE Conference on Computer Vision and Pattern Recognition · 2009 · 61.8K citations
DeCAF: A Deep Convolutional Activation Feature for Generic Visual Recognition
Jeff Donahue, Yangqing Jia, Oriol Vinyals et al. · arXiv (Cornell University) · 2013 · 3.6K citations · Full text
Learning and Transferring Mid-level Image Representations Using Convolutional Neural Networks
Maxime Oquab, Léon Bottou, Ivan Laptev et al. · 2014 · 3.2K citations
Transition-Aware Human Activity Recognition Using Smartphones
Jorge-L. Reyes-Ortiz, Luca Oneto, Albert Samà et al. · Neurocomputing · 2015 · 729 citations