A unified model for human activity recognition using spatial distribution of gradients and difference of Gaussian kernel

Vishwakarma, Dinesh Kumar; Dhiman, Chhavi

doi:10.1007/s00371-018-1560-4

Cited by 58 publications

(13 citation statements)

References 61 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…Ronao et al [22] used smartphones to collect the sensing data of accelerometer and gyroscope, proposed a HAR method based on a two-stage continuous hidden Markov model, and achieved overall recognition accuracy of 91.76%. Vishwakarma et al [23] used the entropy-based texture segmentation method to extract the human silhouette, and proposed a hybrid technology to describe the human activities in the video sequence. Marinho et al [24] proposed a new feature selection method, which uses machine learning technology to realize HAR.…”

Section: A Traditional Machine Learning Methodsmentioning

confidence: 99%

Human Activity Recognition Based on Multichannel Convolutional Neural Network With Data Augmentation

2022

View full text Add to dashboard Cite

In view of the excellent portability and privacy protection of wearable sensor devices, human activity recognition (HAR) of wearable devices has increased applications in human-computer interaction, health care, etc. Therefore, it is necessary to recognize various human activities accurately and efficiently. In this paper, we propose a multi-channel convolutional neural network with data augmentation for HAR, denoted AMC-CNN. First, the sliding windows in time series are used to construct the feature window, and then the feature window is augmented by data transformation and data addition. Through the horizontal connection of the augmented feature windows, the reconstructed feature samples are obtained. Second, a relatively lightweight multi-channel convolutional neural network is designed. In order to mine deep correlations between sensor data more efficiently, we use reasonable convolution kernel settings to compute multi-channel convolutions and extract parallel multi-scale features, and then train the model to recognize human activities. Finally, experiments are carried out on WISDM and MHEALTH datasets to verify the effectiveness of the data augmentation. Moreover, the other three models are adopted for comparative experiments. The evaluation indexes show that the method proposed in this paper has a better recognition effect, which verifies that the AMC-CNN model is effective in both single-sensor and multi-sensor HAR.INDEX TERMS Multi-channel convolutional neural network, data augmentation, human activity recognition, wearable devices.

show abstract

Section: A Traditional Machine Learning Methodsmentioning

confidence: 99%

Human Activity Recognition Based on Multichannel Convolutional Neural Network With Data Augmentation

2022

View full text Add to dashboard Cite

show abstract

“…Hence, we implement a similar technique to our feature extraction method. Vishwakarma et al (Vishwakarma & Singh, 2016;Vishwakama & Dhiman, 2019;Vishwakarma & Kapoor, 2015b) use AESI images, which is a form of silhouette images as a discriminative feature of human shape. Some researchers (Dollar et al, 2006;Efros et al, 2003;Laptev, 2005) incorporate motion information of the video by using spatio-temporal descriptors that effectively capture the characteristics of action.…”

Section: Related Methodsmentioning

confidence: 99%

Section: Related Methodsmentioning

confidence: 99%

A sparse coded composite descriptor for human activity recognition

et al. 2021

Self Cite

View full text Add to dashboard Cite

This paper proposes a novel algorithm for computing discriminative descriptors named as a sparse coded composite descriptor (SCCD) for robust human activity recognition. The proposed method blends the state-of-the-art handcrafted features and the discriminative nature of the sparse representation of visual information. The human activity is firstly modelled using any handcrafted feature, and then the sparse codes computed on a discriminative sparse dictionary of these features are embedded to provide discrimination in the feature set. Finally, a support vector machine (SVM) is trained using the proposed SCCDs to perform classification of different human activities. A new feature named as differential motion descriptor (DMD) is also proposed to extract the motion as well as spatial information from an activity video. The simulation results reveal that in comparison with the handcrafted feature, the corresponding SCCD improves the recognition accuracy significantly. The proposed method is compared with state-of-the-art methods on KTH, Ballet, UCF50, and HMDB51 datasets and the proposed methodology of composite features outperforms these methods in terms of recognition accuracy.

show abstract

“…where high pass filter is used to suppose the edge lines and low pass filter used to smooth the image. But they do not provide more values to palmprint feature extraction except the frequency domain [22,23] .…”

Section: A Difference Of Gaussian Filter (Dog)mentioning

confidence: 99%

A Comparative Analysis of Different Feature Extraction Techniques for Palm-print Images

Pratyasha

Swarnkar

Padhy

2020

Preprint

View full text Add to dashboard Cite

In this advanced decade, automatic identification of individuals is a significant achievement due to the high demand of security system. Hence, individual recognition using biometrics data is leading in the field of image processing. Although biometrics data analysis using thumb impression and finger-prints are very popular since many years, sometimes it leads to false acceptance and rejection if any physical change occurs in the finger ridges. There may be a high risk of hacking the biometrics data which is now a big challenge for cyber security employees. This paper captures the palm-print images of individuals as referred biometrics data for individual recognition. The research work is based on one of the prior issue that is feature extraction to extract the features of palm-print image such as principle lines, textures, ridges and pores etc. For this, some of the feature extraction techniques such as Derivatives of Gaussian filter (DoG), Discrete Cosine Transform (DCT), Fast Fourier Transform (FFT) and competitive coding. Two types of filters: Gaussian Filter and Gabor filter are combined with each of the feature extraction scheme for the matching of sampled image with testing image. In the result, the error rates of each of the feature extraction algorithms are compared to recognize the palm image of two different individuals.

show abstract

A unified model for human activity recognition using spatial distribution of gradients and difference of Gaussian kernel

Cited by 58 publications

References 61 publications

Human Activity Recognition Based on Multichannel Convolutional Neural Network With Data Augmentation

Human Activity Recognition Based on Multichannel Convolutional Neural Network With Data Augmentation

A sparse coded composite descriptor for human activity recognition

A Comparative Analysis of Different Feature Extraction Techniques for Palm-print Images

Contact Info

Product

Resources

About