A reconfigurable interconnected filter for face recognition based on convolution neural network

Dawwd, Shefa A.; Mahmood, Basil Sh.

doi:10.1109/idt.2009.5404141

Cited by 9 publications

(7 citation statements)

References 4 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…There are several benefits to this design approach. It first efficiently lowers data dimensionality and network complexity [1,11]. In addition to helping with feature extraction, weight sharing may be used instead of weight removal, which helps make effective use of the machine's memory.…”

Section: Foundational Context Of the Convolutional Neural Network(cnn)mentioning

confidence: 99%

Gender Classification Using a Convolutional Neural Network (CNN)

Cherukuri,

Nookala,

Balabhadra

et al. 2024

Preprint

View full text Add to dashboard Cite

The purpose of this paper is to demonstrate an innovative convolutional neural network (also known as CNN) methodology for real-time categorization of gender via face photos. The suggested CNN architecture boasts much reduced computational complexity than the current methodologies used in pattern recognition applications. By combining convolutional and subsampling layers, the overall processing layer count is minimised to four. Notably, using cross-correlation versus standard convolution tends to alleviate computing strain. The association is programmed using extended worldwide acquisition frequencies and a second-order backpropagation learning algorithmic framework. The demonstrated CNN approach has been examined using two freely downloadable facial statistics, SUMS and AT&T, with classification accuracies of 99.38% and 98.75%, respectively. Furthermore, the neural network's algorithm demonstrates exceptional efficiency by analysing and categorising a 32 by 32-pixel face picture in just 0.27 milliseconds, resulting in an outstanding consumption of over 3700 images per second. The successful performance of the proposed CNN methodology is further demonstrated by its speedy convergence throughout the training process, which requires less than 20 epochs. The results were produced to showcase the suggested CNN's outstanding accuracy in accurately categorising data, launching it as a realistic and effective real-time identification of the gender system.

show abstract

Section: Foundational Context Of the Convolutional Neural Network(cnn)mentioning

confidence: 99%

Gender Classification Using a Convolutional Neural Network (CNN)

Cherukuri,

Nookala,

Balabhadra

et al. 2024

Preprint

View full text Add to dashboard Cite

show abstract

“…Like in the previous section we also have 4 papers (Dawwd and Mahmood, 2009;Ara et al, 2017;Liu et al, 2019;Zeng et al, 2018a) that did not disclose the training datasets. The authors did not explicitly state why they chose to leave out the datasets, however as they are the same four papers we can assume that the authors did it for the same reasons as before.…”

Section: Assessment Of Q3: What Type Of Cnn Model Is Most Commonly Us...mentioning

confidence: 99%

“…Finally, there are some additional face recognition subjects that were discovered, the first of which is the use of hardware components like those identified in (Dawwd and Mahmood, 2009) where the authors present a reconfigurable hardware model for CNN which is capable of emulating the complex structure of CNN with exploitation of a small chip area by using the property of reconfiguration for the task of face recognition. (Bong et al, 2018) present a low-power CNN face recognition system for user authentication in smart devices.…”

Section: 60%mentioning

confidence: 99%

Convolutional Neural Networks for Face Recognition: A Systematic Literature Review

Barcic¹,

Grd²,

Tomičić³

2023

Preprint

View full text Add to dashboard Cite

This paper presents a comprehensive overview of Convolutional Neural Networks (CNNs) in the context of face recognition. By analyzing 150 research papers, we investigate major publication channels, temporal trends, prevalent CNN models, primary data sets, accuracy levels, primary research focuses, and future prospects for improving CNN-based facial recognition. A major focus is placed on identifying prevalent CNN architectures, techniques used for facial recognition and shedding light on the evolving landscape of CNN designs. Furthermore, we examine the datasets used for training and testing CNNs, and evaluate the accuracy levels achieved by these models. Lastly, we discuss future directions for enhancing CNN-based facial recognition, including addressing bias and fairness, improving robustness to environmental variations, privacy preservation, and exploring transfer learning and multimodal fusion. This paper serves as a valuable resource, summarizing major trends in CNN-based face recognition. It provides insights for researchers and practitioners, guiding future advancements in this rapidly evolving field.

show abstract

“…The next steps of the method are iterative and are performed − times, where is the budget on the number of architecture simulations. The search space is narrowed down from the search history and a new configuration for the next simulation step is suggested using equations presented in Section 3.3 (lines 4,5,6,7,8,9). Once all the iterations have been completed, the optimal architecture configuration set which reaches the minimum EDP is selected (line 10).…”

Section: Hyperopt-gem5-aladdin Frameworkmentioning

confidence: 99%

“…As an example, an architecture may handle a single input feature map and a single output feature map ( _ = 1 and _ = 1), one neuron of each output feature map ( _ = 1 and We evaluated three common workloads. LeNet-5 [14], the most famous handwriting recognition model, FR [9] implementing a face recognition model and HG [15] used to recognize hand gestures of human. In this experiment, we used a non-coherent interface model, it has a private scratchpad memory for local storage and uses DMA to request data from the main memory.…”

Section: Cnn Accelerator In a Socmentioning

confidence: 99%

Design Space Exploration of Heterogeneous-Accelerator SoCs with Hyperparameter Optimization

Cong

Charot

2021

Proceedings of the 26th Asia and South Pacific Design Automation Conference

View full text Add to dashboard Cite

Modern SoC systems consist of general-purpose processor cores augmented with large numbers of specialized accelerators. Building such systems requires a design flow allowing the design space to be explored at the system level with an appropriate strategy. In this paper, we describe a methodology allowing to explore the design space of power-performance heterogeneous SoCs by combining an architecture simulator (gem5-Aladdin) and a hyperparameter optimization method (Hyperopt). This methodology allows different types of parallelism with loop unrolling strategies and memory coherency interfaces to be swept. The flow has been applied to a convolutional neural network algorithm. We show that the most energy efficient architecture achieves a 2x to 4x improvement in energy-delay-product compared to an architecture without parallelism. Furthermore, the obtained solution is more efficient than commonly implemented architectures (Systolic, 2D-mapping, and Tiling). We also applied the methodology to find the optimal architecture including its coherency interface for a complex SoC made up of six accelerated-workloads. We show that a hybrid interface appears to be the most efficient; it reaches 22% and 12% improvement in energy-delay-product compared to just only using non-coherent and only LLC-coherent models, respectively.

show abstract

A reconfigurable interconnected filter for face recognition based on convolution neural network

Cited by 9 publications

References 4 publications

Gender Classification Using a Convolutional Neural Network (CNN)

Gender Classification Using a Convolutional Neural Network (CNN)

Convolutional Neural Networks for Face Recognition: A Systematic Literature Review

Design Space Exploration of Heterogeneous-Accelerator SoCs with Hyperparameter Optimization

Contact Info

Product

Resources

About