Journal
Scientific and Technical Journal of Information Technologies, Mechanics and Optics
UDK
Issue:3 (139)
Download PDF0 Kbyte
Improving sign language processing via few-shot machine learning
Annotation
Improving the efficiency of communication of deaf and hard of hearing people by processing sign language using artificial intelligence is an important task both socially and technologically. One of the ways to solve this problem is a fairly cheap and accessible marker method. The method is based on the registration of electromyographic (EMG) muscle signals using bracelets worn on the arm. To improve the quality of recognition of gestures recorded by the marker method, a modification of the marker method is proposed — duplication of EMG sensors in combination with a low-frame machine learning approach. We experimentally study the possibilities of improving the quality of processing of sign language by duplicating EMG sensors as well as by reducing the volume of the dataset required for training machine learning tools. In the latter case, we compare several technologies of the few-shot approach. Our experiments show that training with few-shot neural nets on 56k samples we can achieve better results than training on random forest with 160k samples. The use of a minimum number of sensors in combination with few-shot signal processing techniques provides the possibility of organizing quick and cost-effective interaction with people with hearing and speech disabilities.
Keywords
Постоянный URL
Articles in current issue
- Methods for audiovisual recognition of people in masks
- Influence of the ratio of the intensities of the reference and object waves on the intensity distribution in the holographic field
- High-precision fiber-optic temperature sensor based on Fabry-Perot interferometer with reflective thin-film multilayer structures
- Optical system design method for the concentration of radiation from a high-power LED
- Detection of potholes on road surfaces using photogrammetry and remote sensing methods
- Adaptive control of nonlinear plant with unmatched parametric uncertainties and input saturation
- Application of failure detection methods to detect information attacks on the control system
- DC motor fault detection and isolation scheme with the use of directional residual set
- Synthesis and implementation of λ-approach of slide control in heat-consumption system
- Photocatalytic properties of Ag-AgBr nanostructures in ion-exchanged layers of bromide sodium-zinc-aluminosilicate glass matrix
- Mechanization of pomset languages in the Coq proof assistant for the specification of weak memory models
- Cloud-based intelligent monitoring system to implement mask violation detection and alert simulation
- Efficient incremental hash chain with probabilistic filter-based method to updateblockchain light nodes
- Method for generating masks on face images and systems for their recognition
- Improving sign language processing via few-shot machine learning
- Quantum-probabilistic SVD: complex-valued factorization of matrix data
- Modelling of basic Indonesian Sign Language translator based on Raspberry Pi technology
- A method of multimodal machine sign language translation for natural human-computer interaction
- Web app for quick evaluation of subjective answers using natural language processing
- Substantiation of construction and evaluation ways of the application efficiency for spatially distributed system of information sensors to provide environment monitoring
- Monitoring of infiltration processes in hydraulic structures using distributed acoustic sensing technology
- Slotted waveguide antenna design for maritime radar system