Motion primitive-based human activity recognition using a bag-of-features approach

Mi, Zhang; Sawchuk, Alexander A.

doi:10.1145/2110363.2110433

Cited by 104 publications

(77 citation statements)

References 14 publications

Supporting

Mentioning

Contrasting

Unclassified

Order By: Relevance

“…Some sensor-based human activity recognition methods employ the codebook approach by making different modifications [22,[27][28][29][30] (Codebook-based methods have also been developed in the fields of sequence/time series classification [25,26,31,32]. However, these are not discussed in this paper, in order to focus on sensor-based human activity recognition.).…”

Section: Related Workmentioning

confidence: 99%

“…For clear discussion, a feature of a subsequence is termed as a local "feature", while the term "feature" is used for a sequence and derived by collecting local features of subsequences. In [28], targeting accelerometer and gyroscope sequences, codewords are extracted by representing subsequences with local features that indicate physical parameters of human motion, such as movement intensities, motion magnitudes along the vertical/heading direction, and the correlation of accelerations between the gravity and heading directions. In [29], for ElectroEncephaloGram (EEG) and ElectroCardioGraphic (ECG) sequences, local features of subsequences are extracted as approximation coefficients obtained by Discrete Wavelet Transform (DWT).…”

Section: Related Workmentioning

confidence: 99%

“…In other words, the vector of raw values in a subsequence is used as its local feature. Hence, our method does not rely on problem-specific local features like the ones in [28,29]. Moreover, in contrast to very rough local features (means and variances) [22,27,30], the detailed information of a subsequence is maintained by its raw values.…”

Section: Related Workmentioning

confidence: 99%

See 2 more Smart Citations

On the Generality of Codebook Approach for Sensor-Based Human Activity Recognition

Shirahama

Grzegorzek

2017

Electronics

View full text Add to dashboard Cite

Abstract:With the recent spread of mobile devices equipped with different sensors, it is possible to continuously recognise and monitor activities in daily life. This sensor-based human activity recognition is formulated as sequence classification to categorise sequences of sensor values into appropriate activity classes. One crucial problem is how to model features that can precisely represent characteristics of each sequence and lead to accurate recognition. It is laborious and/or difficult to hand-craft such features based on prior knowledge and manual investigation about sensor data. To overcome this, we focus on a feature learning approach that extracts useful features from a large amount of data. In particular, we adopt a simple but effective one, called codebook approach, which groups numerous subsequences collected from sequences into clusters. Each cluster centre is called a codeword and represents a statistically distinctive subsequence. Then, a sequence is encoded as a feature expressing the distribution of codewords. The extensive experiments on different recognition tasks for physical, mental and eye-based activities validate the effectiveness, generality and usability of the codebook approach.

show abstract

Section: Related Workmentioning

confidence: 99%

Section: Related Workmentioning

confidence: 99%

See 1 more Smart Citation

On the Generality of Codebook Approach for Sensor-Based Human Activity Recognition

Shirahama

Grzegorzek

2017

Electronics

View full text Add to dashboard Cite

show abstract

“…Mostly, the necessary variability in patterns is captured via Dynamic Time Warping (DTW), allowing the signal to adapt to a pattern [33,34]. Another possibility is decomposing activities in simple motion primitives and applying a bag-of-features model [35].…”

Section: Introductionmentioning

confidence: 99%

Recognition of Physical Activities from a Single Arm-Worn Accelerometer: A Multiway Approach

et al. 2018

View full text Add to dashboard Cite

Abstract:In current clinical practice, functional limitations due to chronic musculoskeletal diseases are still being assessed subjectively, e.g., using questionnaires and function scores. Performance-based methods, on the other hand, offer objective insights. Hence, they recently attracted more interest as an additional source of information. This work offers a step towards the shift to performance-based methods by recognizing standardized activities from continuous readings using a single accelerometer mounted on a patient's arm. The proposed procedure consists of two steps. Firstly, activities are segmented, including rejection of non-informative segments. Secondly, the segments are associated to predefined activities using a multiway pattern matching approach based on higher order discriminant analysis (HODA). The two steps are combined into a multi-layered framework. Experiments on data recorded from 39 patients with spondyloarthritis show results with a classification accuracy of 94.34% when perfect segmentation is assumed. Automatic segmentation has 89.32% overlap with this ideal scenario. However, combining both drops performance to 62.34% due to several badly-recognized subjects. Still, these results are shown to significantly outperform a more traditional pattern matching approach. Overall, the work indicates promising viability of the technique to automate recognition and, through future work, assessment, of functional capacity.

show abstract

“…BOVW originated from the Bag of Words (BOW) model which was initially used in document processing, but has recently been introduced into the imaging community [5]. Since then, several derivatives and refinements have been proposed to adapt to different applications (i.e.…”

Section: Introductionmentioning

confidence: 99%

Nearest-Neighbors Based Weighted Method for the BOVW Applied to Image Classification

Xu¹,

Sun²,

Lu³

et al. 2015

Journal of Electrical Engineering and Technology

View full text Add to dashboard Cite

-This paper presents a new Nearest-Neighbors based weighted representation for images and weighted K-Nearest-Neighbors (WKNN) classifier to improve the precision of image classification using the Bag of Visual Words (BOVW) based models. Scale-invariant feature transform (SIFT) features are firstly extracted from images. Then, the K-means++ algorithm is adopted in place of the conventional K-means algorithm to generate a more effective visual dictionary. Furthermore, the histogram of visual words becomes more expressive by utilizing the proposed weighted vector quantization (WVQ). Finally, WKNN classifier is applied to enhance the properties of the classification task between images in which similar levels of background noise are present. Average precision and absolute change degree are calculated to assess the classification performance and the stability of K-means++ algorithm, respectively. Experimental results on three diverse datasets: Caltech-101, Caltech-256 and PASCAL VOC 2011 show that the proposed WVQ method and WKNN method further improve the performance of classification.

show abstract

Motion primitive-based human activity recognition using a bag-of-features approach

Cited by 104 publications

References 14 publications

On the Generality of Codebook Approach for Sensor-Based Human Activity Recognition

On the Generality of Codebook Approach for Sensor-Based Human Activity Recognition

Recognition of Physical Activities from a Single Arm-Worn Accelerometer: A Multiway Approach

Nearest-Neighbors Based Weighted Method for the BOVW Applied to Image Classification

Contact Info

Product

Resources

About