Abstract:Deep convolutional neural networks (DCNNs) are currently popular in human activity recognition (HAR) applications. However, in the face of modern artificial intelligence sensor-based games, many research achievements cannot be practically applied on portable devices (i.e., smart phone, VR/AR). DCNNs are typically resource-intensive and too large to be deployed on portable devices, and thus, this limits the practical application of complex activity detection. In addition, since portable devices do not possess high-performance graphic processing units, there is hardly any improvement in Action Game (ACT) experience. Besides, in order to deal with multi-sensor collaboration, all previous HAR models typically treated the representations from different sensor signal sources equally. However, distinct types of activities should adopt different fusion strategies. In this paper, a novel scheme is proposed. This scheme is used to train 2-bit CNNs with weights and activations constrained to {-0.5, 0, 0.5}. It takes into account the correlation between different sensor signal sources and the activity types. This model, which we refer to as DFTerNet, aims at producing a more reliable inference and better trade-offs for practical applications. It is known that quantization of weights and activations can substantially reduce memory size and use more efficient bitwise operations to replace floating or matrix operations to achieve much faster calculation and lower power consumption. Our basic idea is to exploit quantization of weights and activations directly in pre-trained filter banks and adopt dynamic fusion strategies for different activity types. Experiments demonstrate that by using a dynamic fusion strategy, it is possible to exceed the baseline model performance by up to ~5% on activity recognition data sets, such as the OPPORTUNITY and PAMAP2 data sets. Using the quantization method proposed, we were able to achieve performances closer to that of the full-precision counterpart. These results were also verified using the UniMiB-SHAR data set. In addition, the proposed method can achieve ~9x acceleration on CPUs and ~11x memory saving.

A Collaborative Compression Scheme for Fast Activity Recognition on Mobile Devices Via Global Compression Ratio Decision

Cloud-Edge Collaborative Human Activity Recognition System Design and Implementation.

MCMC: Multi-Constrained Model Compression Via One-Stage Envelope Reinforcement Learning.

Real-time Human Activity Recognition Using Conditionally Parametrized Convolutions on Mobile and Wearable Devices

On-Demand Deep Model Compression for Mobile Devices

Deep Convolutional Networks With Tunable Speed–Accuracy Tradeoff for Human Activity Recognition Using Wearables

Deep Model Compression for Mobile Platforms: A Survey

A Fast And Robust Deep Convolutional Neural Networks For Complex Human Activity Recognition Using Smartphone

A Human Activity Recognition Method Based on Lightweight Feature Extraction Combined with Pruned and Quantized CNN for Wearable Device

An Efficient Diverse-Branch Convolution Scheme for Sensor-Based Human Activity Recognition

Towards Compact CNNs via Collaborative Compression

A CNN Compression Method Via Dynamic Channel Ranking Strategy

An automatic network structure search via channel pruning for accelerating human activity inference on mobile devices

Holistic CNN Compression Via Low-Rank Decomposition with Knowledge Transfer.

Layer-Wise Training Convolutional Neural Networks With Smaller Filters for Human Activity Recognition Using Wearable Sensors

A multi-scale feature extraction fusion model for human activity recognition

DFTerNet: Towards 2-Bit Dynamic Fusion Networks for Accurate Human Activity Recognition.

Multiscale Deep Feature Learning for Human Activity Recognition Using Wearable Sensors

ANS: Assimilating Near Similarity at High Accuracy for Significant Deduction of CNN Storage and Computation

GrapHAR: A Lightweight Human Activity Recognition Model by Exploring the Sub-carrier Correlations

Wearable Sensor-Based Residual Multifeature Fusion Shrinkage Networks for Human Activity Recognition