An enhanced method for human action recognition

Moussa, Mona M.; Hamayed, Elsayed; Fayek, Magda B.; Nemr, Heba A. El

doi:10.1016/j.jare.2013.11.007

Cited by 41 publications

(15 citation statements)

References 17 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…Vreal [f1 = A02; f2 = A22; f3 = A33; f4 = A24]; (10) Vmagnitude [f1 = A00; f2 = A02; f3 = A15; f4 = A55]: (11) When the Zernike moment feature vectors Vreal and Vmagnitude, as defined in (7) and (8), are applied, we only need to consider the real part or magnitude of A nm : TX' , ' 5 Y = 7 ' − ' 5 8 [5 8 (12) Where i and j denote any pair of human actions of training and testing of video frames respectively. From the result of accuracy recognition in Table. 4.…”

Section: Resultsmentioning

confidence: 99%

“…Khan et al [14], and Moussa et al [10], propose methods of action recognition with performing quite high recognition rate, because in [14], they used whole parts of seven Hu invariant moments, and in [10] used normalization after feature detection of interest points, this means more cost in terms of computation time, their method, time is 14.446, although the accuracy is high. While our method use four feature vectors, so our method performs faster than [10] even though calculating more feature vectors. The difficulties of other similar related methods as in [19] are to classify between jogging and running, they used SIFT descriptor for identity representation and ASIFT descriptor for action representation in expensive computing, their method of action accuracy is 81.7 and for identity accuracy is 58.8.…”

Section: Resultsmentioning

confidence: 99%

“…Finally Ballan et al [9] developed an efficient descriptor, describing the spatial and temporal components and creating separated histograms of 3D gradient orientations. Moussa et al [10] proposed an enhanced method for human action recognition by using fine tuning to reduce the number of interest point detectors that produced from SIFT algorithm, they get high accuracy of action recognition. Khawlah et al [11] proposed identity as watermark with SIFT descriptors.…”

Section: Related Workmentioning

confidence: 99%

“…Based on the author's method [10], we develop the watermarked method, which the watermark can be embedded into each frame of the video. Digital watermarking [20] is the process of embedding or hiding digital information called watermark into a multimedia product, and then the embedded data can later be extracted or detected from the watermarked product.…”

Section: Identify the Human By Using A Watermarkmentioning

confidence: 99%

See 3 more Smart Citations

On the Development Robust and Fast Algorithm of Action and Identity Recognition

Alhamzah¹

2015

AJNC

View full text Add to dashboard Cite

Human action recognition and surveillance applications are playing a key important in the present days and took an increasing interest in modern. Since most previous methods strictly limited to action classification in different scenarios and not take attention to human identity that makes an action at the same time. We present a novel and fast algorithm to recognize action and identity in a single framework. We assumed one person makes one action in a video. To identify and training the owner of the video to the classifier, we proposed the watermark embedded as 2-D wavelet transform as binary image, which is contains identity information in the training video. We used these wavelet coefficients as identity descriptors. To represent feature motion representation, we used motion energy image (MEI) and motion history image (MHI) as temporal template of the human actions and Zernike moments to extract shape features of the action from MEI and MHI. In this research, a set of Zernike moment based feature vectors is proposed for human action recognition, which is capture the global properties of an object rather than the local ones. We have composed two different feature vectors by evaluating the variance values of lower order Zernike moments in the four-dimensional Zernike moment space with encouraging experimental results. It has discriminative information that is suitable for classification, especially on related actions, such as running and jogging, that is most previous researches fail to classify them even human vision HVS. Nearest neighbor classifier is used for action and identity categorization. The result of these experiments suggests that this method has a high recognition rate in both action and identity accuracy on KTH data sets.

show abstract

Section: Resultsmentioning

confidence: 99%

Section: Resultsmentioning

confidence: 99%

Section: Related Workmentioning

confidence: 99%

Section: Identify the Human By Using A Watermarkmentioning

confidence: 99%

See 2 more Smart Citations

On the Development Robust and Fast Algorithm of Action and Identity Recognition

Alhamzah¹

2015

AJNC

View full text Add to dashboard Cite

show abstract

“…In [12], BoVW technique is used for human action recognition. The method’s training phase consists of five steps.…”

Section: Walking Pattern Detectionmentioning

confidence: 99%

RGB-D camera based walking pattern recognition by support vector machines for a smart rollator

Zhang

2017

Int J Intell Robot Appl

View full text Add to dashboard Cite

This paper presents a walking pattern detection method for a smart rollator. The method detects the rollator user’s lower extremities from the depth data of an RGB-D camera. It then segments the 3D point data of the lower extremities into the leg and foot data points, from which a skeletal system with 6 skeletal points and 4 rods is extracted and used to represent a walking gait. A gait feature, comprising the parameters of the gait shape and gait motion, is then constructed to describe a walking state. K-means clustering is employed to cluster all gait features obtained from a number of walking videos into 6 key gait features. Using these key gait features, a walking video sequence is modeled as a Markov chain. The stationary distribution of the Markov chain represents the walking pattern. Three Support Vector Machines (SVMs) are trained for walking pattern detection. Each SVM detects one of the three walking patterns. Experimental results demonstrate that the proposed method has a better performance in detecting walking patterns than seven existing methods.

show abstract

An Approach Towards Action Recognition Using Part Based Hierarchical Fusion

Agarwal

Sen

2020

Advances in Visual Computing

View full text Add to dashboard Cite

An enhanced method for human action recognition

Cited by 41 publications

References 17 publications

On the Development Robust and Fast Algorithm of Action and Identity Recognition

On the Development Robust and Fast Algorithm of Action and Identity Recognition

RGB-D camera based walking pattern recognition by support vector machines for a smart rollator

An Approach Towards Action Recognition Using Part Based Hierarchical Fusion

Contact Info

Product

Resources

About