A multi-task fully deep convolutional neural network for contactless fingerprint minutiae extraction

Zhang, Zhao; Liu, Shuxin; Liu, Manhua

doi:10.1016/j.patcog.2021.108189

Cited by 26 publications

(14 citation statements)

References 24 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…Multi-task learning has proved its potential to develop future intelligent system, such as obstacle avoidance and object tracking in autonomous driving [96]. In conventional software implemented neural network, multi-task learning is achieved by introducing a new loss function, which jointly combine two target tasks [97]. When it comes to hardware network, neuron devices will be capable for multi-task learning as long as appending more input ports to the device, jointly modulating the membrane potential.…”

Section: Snnmentioning

confidence: 99%

Emerging functions of two-dimensional materials in memristive neurons

Hong,

Liu,

et al. 2024

J. Phys. Mater.

View full text Add to dashboard Cite

Neuromorphic computing (NC), considered as a promising candidate for future computer architecture, can facilitate more biomimetic intelligence while reducing energy consumption. Neuron is one of the critical building blocks of NC systems. Researchers have been engaged in promoting neuron devices with better electrical properties and more biomimetic functions. Two-dimensional (2D) materials, with ultrathin layers, diverse band structures, featuring excellent electronic properties and various sensing abilities, are promised to realize these requirements. Here, the progress of artificial neurons brought by 2D materials is reviewed, from the perspective of electrical performance of neuron devices, from stability, tunability to power consumption and on/off ratio. Rose up to system-level applications, algorithms and hardware implementation of spiking neural network, stochastic neural network and artificial perception system based on 2D materials are reviewed. 2D materials not only facilitate the realization of NC systems but also increase the integration density. Finally, current challenges and perspectives on developing 2D material-based neurons and NC systems are systematically analyzed, from the bottom 2D materials fabrication to novel neural devices, more brain-like computational algorithms and systems.

show abstract

Section: Snnmentioning

confidence: 99%

Emerging functions of two-dimensional materials in memristive neurons

Hong,

Liu,

et al. 2024

J. Phys. Mater.

View full text Add to dashboard Cite

show abstract

“…Usually in Indian Classical Music a singer sticks to a particular Sa frequency throughout his performances and sometimes throughout his career with exceptions. By extracting features and training a Neural Network with performances of various singers it may be possible to identify the most appropriate Sa frequency for a given performance of a musician 34,43,44,49,50 …”

Section: Introductionmentioning

confidence: 99%

“…By extracting features and training a Neural Network with performances of various singers it may be possible to identify the most appropriate Sa frequency for a given performance of a musician. 34,43,44,49,50 Various methods of raga identification have been tried by researchers. 24,25 Identifying pitches and the Shadja would help in identifying the note names and the scale of a raga.…”

mentioning

confidence: 99%

An improved sub‐harmonic to harmonic ratio method for pitch estimation and Shadja detection

Kumaraswamy¹,

Poonacha

2023

Concurrency and Computation

View full text Add to dashboard Cite

In this paper, we have presented an improved Sub-Harmonic to Harmonic ratio (SHR) algorithm using the Genetic algorithm (GA) for pitch estimation of audio recordings of various ragas. Then we study the problem of Shadja identification using machine learning with the help of feature extraction and classification. The extraction of features from the raga signal is done with the help of statistical analysis of pitch estimation and these features are classified using a Neural Network (NN). Here, the training of NN is accomplished by Gray Wolf Optimization (GWO) algorithm for, determining the weights. Performance of the proposed Sa detection algorithm is analyzed by comparing the developed NN using Gray Wolf Optimization (GWO) models with the conventional models such as Levenberg Marquardt based NN (LM-NN), Gradient Descent based (GD-NN), Particle Swarm Optimization based NN (PSO-NN) and FireFly based NN (FF-NN) in terms of positive and negative performance measures.

show abstract

“…The controller is designed through offline data analysis and deployed for online use. This offline-training-onlinepractice (OTOP) mode offers more adaptability and accuracy of control, and has been applied in various fields (Nawfel et al, 2021;Zhang et al, 2021aZhang et al, , 2021bWang et al, 2022aWang et al, , 2022b.…”

Section: Introductionmentioning

confidence: 99%

“…In this paper, the OTOP mode is incorporated into model-free reinforcement learning (RL). Recent studies show that RL can cope with high-dimensional stochastic problems (Vanvuchelen et al, 2020;Zhang et al, 2021aZhang et al, , 2021bHe et al, 2021). During the training process, the agent is not told what to do, but instead must discover which actions yield the most reward by trying them (Sutton and Barto, 2018).…”

Section: Introductionmentioning

confidence: 99%