On prediction error compressive sensing image reconstruction for face recognition

Biswas, Suparna; Sil, Jaya; Maity, Santi P.

doi:10.1016/j.compeleceng.2017.11.009

Cited by 7 publications

(4 citation statements)

References 7 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…This information is then used to guide compression. The method [41] examines the recognition of a remote face recognition (FR) scheme in the framework of compact image recognition (CS). Error images are obtained as the difference between the original and predicted images, using various predictors such as Autoencoder that provide compact measurements.…”

Section: -Related Workmentioning

confidence: 99%

WITHDRAWN: A new method of facial image compression based on meta-heuristic algorithms with variable bit budget allocation

khodadadi

Ardeshir

Grailu

2022

Preprint

View full text Add to dashboard Cite

One of the important research areas in imaging is the formation of images, which plays an important role in many different applications, including surveillance, control, and security affairs. On the other hand, high spatial resolution is one of the most important factors for increasing image quality, but it increases the amount of storage memory. Based on this, it is expected that the processing methods will be concentrated in this area. In face recognition systems, one of the existing challenges is maintaining the image recognition rate. These images in the recognition system may affect the efficiency of the system, and as a result, it may be inferred that the percentage of recognition is found. Proposing a method that at least does not reduce detection rates would be very desirable. This article investigates how to compress facial images with high spatial resolution using innovative algorithms to reduce or even increase their accuracy as much as possible. In this article, meta-heuristic algorithms (genetics and gray wolf) are used in a way that they are responsible for identifying the important and similar areas of matching macroblocks in the whole image segmentation. So how arranging the length of the bit string of each block is appropriate to achieve the target estimation to achieve the highest recognition accuracy value, PSNR value, and SSIM value of the image set? In the simulation and evaluation section, the facial images of the CIE and FEI databases have been examined as a selective study, and the recognition efficiency of the images for the conditions without/with compression and the common SPIHT and JPEG coding methods and compared with the proposed method. The simulation results show the significant impact of the proposed methods using meta-heuristic algorithms in increasing the quality of PSNR and SSIM in contrast to the recognition efficiency. According to the proposed method, the larger the value of dividing the blocks, the better the average PSNR and SSIM, assuming the highest recognition accuracy can be reached in both meta-heuristic algorithms. In general, depending on the type of application of the problem, there is a compromise to achieve the highest average PSNR or SSIM, using a genetic algorithm or gray wolf. The gray wolf algorithm, however, reaches its optimal answer much faster than the genetic algorithm.

show abstract

Section: -Related Workmentioning

confidence: 99%

WITHDRAWN: A new method of facial image compression based on meta-heuristic algorithms with variable bit budget allocation

khodadadi

Ardeshir

Grailu

2022

Preprint

View full text Add to dashboard Cite

show abstract

“…In terms of the recognition rates, Ours outperformed the state-ofthe-art methods, including hand-crafted features and deep features techniques. Also, our WSBP was better than Multi-resolution dictionary [30] (82.19%), MNTCDP [21] (96.18%), Local Multiple Patterns [49] (98.00%), or even deep facial features CS [2] (93.99%) by a substantial margin. The remaining algorithms, including EL-LBP [44] (98.27%) and deep feature FDDL + CNN [39] (98%) were comparable with our descriptors, and yet ours prevailed.…”

Section: Evaluation With Gray-scale Imagesmentioning

confidence: 89%

“…The average recognition rates from the different methods were shown in Table 12. Here, we fine-tuned the structuring element B 2 = { (1,5); (2,6)} to obtain the best achievements for WSBP 99.03 Görgel et al [13] 97.50 and CLBP S M(m 1 , μ 2 ). This structuring element made our descriptors more robust against noise and occlusion comparing to other methods.…”

Section: Robustness Against Degraded Imagesmentioning

confidence: 99%

Weighted statistical binary patterns for facial feature representation

2021

View full text Add to dashboard Cite

We present a novel framework for efficient and robust facial feature representation based upon Local Binary Pattern (LBP), called Weighted Statistical Binary Pattern, wherein the descriptors utilize the straight-line topology along with different directions. The input image is initially divided into mean and variance moments. A new variance moment, which contains distinctive facial features, is prepared by extracting root k-th. Then, when Sign and Magnitude components along four different directions using the mean moment are constructed, a weighting approach according to the new variance is applied to each component. Finally, the weighted histograms of Sign and Magnitude components are concatenated to build a novel histogram of Complementary LBP along with different directions. A comprehensive evaluation using six public face datasets suggests that the present framework outperforms the state-of-the-art methods and achieves 98.51% for ORL, 98.72% for YALE, 98.83% for Caltech, 99.52% for AR, 94.78% for FERET, and 99.07% for KDEF in terms of accuracy, respectively. The influence of color spaces and the issue of degraded images are also analyzed with our descriptors. Such a result with theoretical underpinning confirms that our descriptors are robust against noise, illumination variation, diverse facial expressions, and head poses.

show abstract

“…As a result of this information, compression is guided. This method [40] examines the recognition of remote face recognition (FR) schemes within the framework of compact image recognition (CS). Using predictors such as autoencoder that provide compact measurements, error images can be obtained as the difference between original and predicted images.…”

Section: Related Workmentioning

confidence: 99%

Compressing Face Images Using Genetic and Gray Wolf Meta-heuristic Algorithms Based on Variable Bit Allocation

khodadadi

Ardeshir

Grailu

2023

IJE

View full text Add to dashboard Cite

In image processing, compression plays an important role in monitoring, controlling, and securing the process. The spatial resolution is one of the most effective factors in improving the quality of an image; but, it increases the amount of storage memory required. Based on meta-heuristic algorithms, this article presents a compression model for face images with block division and variable bit allocation. Wavelet transform is used to reduce the dimensions of high spatial resolution face images. In order to identify important and similar areas of identical macroblocks, genetic algorithms and gray wolves are used. A bit rate allocation is calculated for each block to achieve the best recognition accuracy, average PSNR, and SSIM. The CIE and FEI databases have been used as case studies. The proposed method has been tested and compared with the accuracy of image recognition under uncompressed conditions and using the common SPIHT and JPEG coding methods. Recognition accuracy increased from 0.18% for 16×16 blocks to 1.97% for 32×32 blocks. Additionally, the gray wolf algorithm is much faster than the genetic algorithm in reaching the optimal answer. Depending on the application type of the problem, the genetic algorithm or the gray wolf may be preferred to achieve the maximum average PSNR or SSIM. At the bit rate of 0.9, the maximum average PSNR for the gray wolf algorithm is 34.92 and the maximum average SSIM for the genetic algorithm is 0.936. Simulation results indicate that the mentioned algorithms increase PSNR and SSIM by stabilizing or increasing recognition accuracy.

show abstract

On prediction error compressive sensing image reconstruction for face recognition

Cited by 7 publications

References 7 publications

WITHDRAWN: A new method of facial image compression based on meta-heuristic algorithms with variable bit budget allocation

WITHDRAWN: A new method of facial image compression based on meta-heuristic algorithms with variable bit budget allocation

Weighted statistical binary patterns for facial feature representation

Compressing Face Images Using Genetic and Gray Wolf Meta-heuristic Algorithms Based on Variable Bit Allocation

Contact Info

Product

Resources

About