Inception inspired CNN-GRU hybrid network for human activity recognition

Dua, Nidhi; Singh, Shiva Nand; Semwal, Vijay Bhaskar; Challa, Sravan Kumar

doi:10.1007/s11042-021-11885-x

Cited by 68 publications

(25 citation statements)

References 59 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…In addition to depth and RGB monitoring systems, using a wearable sensor is another alternative to detect and analyze the human posture; among them, methods such as [62][63][64][65][66][67][68][69] can be mentioned. Although the accuracy of these methods is very promising, they need the voluntary cooperation of the subject, and they are prone to be forgotten.…”

Section: Discussionmentioning

confidence: 99%

On Study of 1D Depth Scans as an Alternative Feature for Human Pose Detection in a Sensor Network

Rasoulidanesh

Payandeh

2022

Journal of Sensors

View full text Add to dashboard Cite

Inspired by the notion of swarm robotics, sensing, and minimalism, in this paper, we study and analyze how a collection of only 1D depth scans can be used as a part of the minimum feature for human body detection and its segmentation in a point cloud. In relation to the traditional approaches which require a complete point cloud model representation for skeleton model reconstruction, our proposed approach offers a lower computation and power consumption, especially in sensor and robotic networks. Our main objective is to investigate if the reduced number of training data through a collection of 1D scans of a subject is related to the rate of recognition and if it can be used to accurately detect the human body and its posture. The method takes advantage of the frequency components of the depth images (here, we refer to it as a 1D scan). To coordinate a collection of these 1D scans obtained through a sensor network, we also proposed a sensor scheduling framework. The framework is evaluated using two stationary depth sensors and a mobile depth sensor. The performance of our method was analyzed through movements and posture details of a subject having two relative orientations with respect to the sensors with two classes of postures, namely, walking and standing. The novelty of the paper can be summarized in 3 main points. Firstly, unlike deep learning methods, our approach would require a smaller dataset for training. Secondly, our case studies show that the method uses very limited training dataset and still can detect the unseen situation and reasonably estimate the orientation and detail of the posture. Finally, we propose an online scheduler to improve the energy efficiency of the network sensor and minimize the number of sensors required for surveillance monitoring by employing a mobile sensor to recover the occluded views of the stationary sensors. We showed that with the training data captured on 1 m from the camera, the algorithm can detect the detailed posture of the subject from 1, 2, 3, and 4 meters away from the sensor during the walking and standing with average accuracy of 93% and for different orientation with respect to the sensor by 71% accuracy.

show abstract

Section: Discussionmentioning

confidence: 99%

On Study of 1D Depth Scans as an Alternative Feature for Human Pose Detection in a Sensor Network

Rasoulidanesh

Payandeh

2022

Journal of Sensors

View full text Add to dashboard Cite

show abstract

“…In order to retain recurrent information of the time series data the modules are integrated in parallel where the LSTM module is connected to an inception time network with additional layers of attention (Abbasimehr and Paki, 2022). Novel FDI techniques are proposed in comparison to six theft cases for data manipulation (Dua et al, 2022). AttenLSTMInception model is a multivariate resolution feature of the time series data.…”

Section: Model's Architecturementioning

confidence: 99%

Novel FDIs-based data manipulation and its detection in smart meters’ electricity theft scenarios

Munawar

Khan²,

Chaudhary³

et al. 2022

Front. Energy Res.

View full text Add to dashboard Cite

Non-technical loss is a serious issue around the globe. Consumers manipulate their smart meter (SM) data to under-report their readings for financial benefit. Various manipulation techniques are used. This paper highlights novel false data injection (FDIs) techniques, which are used to manipulate the smart meter data. These techniques are introduced in comparison to six theft cases. Furthermore, various features are engineered to analyze the variance, complexity, and distribution of the manipulated data. The variance and complexity are created in data distribution when FDIs and theft cases are used to poison SM data, which is investigated through skewness and kurtosis analysis. Furthermore, to tackle the data imbalance issue, the proximity weighted synthetic oversampling (ProWsyn) technique is used. Moreover, a hybrid attentionLSTMInception is introduced, which is an integration of attention layers, LSTM, and inception blocks to tackle data dimensionality, misclassification, and high false positive rate issues. The proposed hybrid model outperforms the traditional theft detectors and achieves an accuracy of 0.95%, precision 0.97%, recall 0.94%, F1 score 0.96%, and area under-the-curve (AUC) score 0.98%.

show abstract

“…Inspired by the breakthrough of CNN based architectures in several computer vision applications, 8‐12,25‐27 researchers have also used CNN for video‐based HAR tasks 11,12 due to its self‐learning capability from raw data. In Reference 11, Khan et al have proposed a framework for HAR for surveillance applications.…”

Section: Literature Reviewmentioning

confidence: 99%

“…Also, it is always challenging to choose the fitting feature descriptor for a specific problem. 24 Inspired by the breakthrough of CNN based architectures in several computer vision applications, [8][9][10][11][12][25][26][27] researchers have also used CNN for video-based HAR tasks 11,12 due to its self-learning capability from raw data. In Reference 24 have proposed a 26-layer deep CNN architecture to classify complex human activities.…”

Section: Literature Reviewmentioning

confidence: 99%

Human activity recognition based on integration of multilayer information of convolutional neural network architecture

Kushwaha

Srivastava

Khare

2022

Concurrency and Computation

View full text Add to dashboard Cite

Human activity recognition (HAR) has gained researcher's interest due to its increasing demand in automated monitoring applications. Development of efficient HAR algorithm is still an open research area due to the challenges like inter and intra-class variations, diversity in lighting conditions, view point changes, and complex object motions. Convolutional neural network (CNN) based methods have achieved significant improvement in HAR. However, CNN implementations have drawback that it require a lot of computational resources due to the use of large number of learnable parameters. To overcome this drawback, we propose a simple and computationally efficient deep CNN architecture using multi-layer information fusion for HAR. In this study, we explore the impact of information fusion at intermediate layers of the network, as each convolutional layer of the network hierarchically extracts information at different level of abstraction of the objects from the video frames. In this work, first we designed a simple and computationally efficient deep CNN architecture and then we introduce a feature fusion strategy to integrate the complementary information of intermediate layers to the layer of the proposed CNN architecture. The proposed architecture is fine-tuned and trained from scratch with raw RGB data. Softmax classifier is used at the last layer of network for activity classification. Benefits of the proposed architecture over standard deep learning architectures is it's high computational efficiency and reduced requirement of computational resources. To prove the effectiveness of the proposed method, we performed several extensive experiments on publically available datasets. The experimental results of the proposed method have demonstrated its superiority over other existing state-of-the-art methods.

show abstract

Inception inspired CNN-GRU hybrid network for human activity recognition

Cited by 68 publications

References 59 publications

On Study of 1D Depth Scans as an Alternative Feature for Human Pose Detection in a Sensor Network

On Study of 1D Depth Scans as an Alternative Feature for Human Pose Detection in a Sensor Network

Novel FDIs-based data manipulation and its detection in smart meters’ electricity theft scenarios

Human activity recognition based on integration of multilayer information of convolutional neural network architecture

Contact Info

Product

Resources

About