Face Recognition in Home Security System Using Tensor Decomposition Based on Radix-(2 × 2) Hierarchical SVD

Kountchev, Roumen; Anwar, Suzan; Kountcheva, Roumiana; Milanova, Mariofanna

doi:10.1007/978-3-319-59259-6_5

Cited by 3 publications

(4 citation statements)

References 8 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…In Table 1 is given the recognition rate for same training and testing images, but using ТHSVD. The table shows that the rate is increased and reached 100% along with changing of the number of training and testing images [16]. The 100% recognition rate follows from the use of ТHSVD instead of 2D SVD.…”

Section: Resultsmentioning

confidence: 94%

“…For the evaluation of the proposed new method were executed some experiments for face recognition, based on the TensorFace method [16]. The method requires a training set of faces which to be compared to new faces.…”

Section: Resultsmentioning

confidence: 99%

“…Each 2D-LUT memory should have 2 2К cells. For example, if K=8, the 2D-LUT should be of 2 16 =64 KB cells, where the values of the products m, n, g, h are saved as 8-bit words.…”

Section:  mentioning

confidence: 99%

See 2 more Smart Citations

Image Sequence Decomposition Based on the Truncated Hierarchical SVD

Kountchev¹,

Kountcheva²

2017

IJMIP

Self Cite

View full text Add to dashboard Cite

In this work is presented new algorithm, called Truncated Hierarchical SVD, which is the advance of the algorithm Radix2×2 Hierarchical SVD, developed earlier by the authors of this paper. The new approach is aimed at the processing of sequences of correlated images, represented as 3 rd -order tensors. It is based on the multiple execution of SVD for elementary tensors (ЕТ) of size 222, which build the tensor of size NNN, when N=2 n . The new approach is compared here to closest famous hierarchical SVD method for ЕТ -the Sequential Unfolding SVD. The comparison proves, that the new algorithm has lower computational complexity. To reduce the computational complexity of the algorithm, here is used "truncation" of tensor components with small weights, based on new adaptive approach. As a result of the "truncation" and of the parallel implementation, the processing of image sequences, represented by 3 rd order tensors, is significantly accelerated. These advantages open new abilities for real-time application in image processing systems.

show abstract

Section: Resultsmentioning

confidence: 94%

Section: Resultsmentioning

confidence: 99%

See 1 more Smart Citation

Image Sequence Decomposition Based on the Truncated Hierarchical SVD

Kountchev¹,

Kountcheva²

2017

IJMIP

Self Cite

View full text Add to dashboard Cite

show abstract

“…The tensor structure provides an excellent method for representing many kinds of highly correlated data such as videos. It is used in many applications as in [ 27 , 28 , 29 ]. Cheng et al [ 8 ] discussed tensor data decomposition and its great influence on dimension reduction.…”

Section: Related Workmentioning

confidence: 99%

Detecting and Locating Passive Video Forgery Based on Low Computational Complexity Third-Order Tensor Representation

2021

View full text Add to dashboard Cite

Great attention is paid to detecting video forgeries nowadays, especially with the widespread sharing of videos over social media and websites. Many video editing software programs are available and perform well in tampering with video contents or even creating fake videos. Forgery affects video integrity and authenticity and has serious implications. For example, digital videos for security and surveillance purposes are used as evidence in courts. In this paper, a newly developed passive video forgery scheme is introduced and discussed. The developed scheme is based on representing highly correlated video data with a low computational complexity third-order tensor tube-fiber mode. An arbitrary number of core tensors is selected to detect and locate two serious types of forgeries which are: insertion and deletion. These tensor data are orthogonally transformed to achieve more data reductions and to provide good features to trace forgery along the whole video. Experimental results and comparisons show the superiority of the proposed scheme with a precision value of up to 99% in detecting and locating both types of attacks for static as well as dynamic videos, quick-moving foreground items (single or multiple), zooming in and zooming out datasets which are rarely tested by previous works. Moreover, the proposed scheme offers a reduction in time and a linear computational complexity. Based on the used computer’s configurations, an average time of 35 s. is needed to detect and locate 40 forged frames out of 300 frames.

show abstract