DeSIRe: Deep Signer-Invariant Representations for Sign Language Recognition

Ferreira, Pedro M.; Pernes, Diogo; Rebelo, Ana; Cardoso, Jaime S.

doi:10.1109/tsmc.2019.2957347

Cited by 10 publications

(9 citation statements)

References 36 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…Without visual information, there is a tendency to miss key cues and make an incorrect judgment. Therefore, acting like a visual chatbot, ViLT considers nonverbal information, such as facial expressions, gestures, and actions [80], [81], to better understand worker behavior and generate more appropriate responses.…”

Section: Multimodal Multitask Foundation Modelmentioning

confidence: 99%

Integrated Inspection on PCB Manufacturing in Cyber–Physical–Social Systems

Wang

Cao

et al. 2023

IEEE Trans. Syst. Man Cybern, Syst.

View full text Add to dashboard Cite

The printed circuit boards (PCBs) industry is one of the fastest-growing industries in recent decades. The PCB manufacturing process is highly complicated and severely affected by social factors, which makes it very important to conduct integrated inspection, assuring and improving the production quality. In this article, we propose an artificial systems, computational experiments, and parallel execution-based integrated inspection method in cyber-physical-social systems (CPSS) to realize smart manufacturing. In this inspection system, rather than simply performing modeling, analysis, and control, we perform descriptive intelligence to construct production processes with limited multimodal information, perform predictive intelligence to conduct defect detection and defect prediction, and perform prescriptive intelligence to achieve defect diagnosis and defect management. In this way, our inspection system could offer a learning and training platform for workers to master professional inspection skills, provide an experimentation and evaluation platform for product defect monitoring and early warnings, and supply guidance about defect management and control to improve manufacturing processes. For technical implementation, we leverage a Transformer-based foundation model to achieve knowledge reasoning and human-computer interaction. As a result, we provide an innovative solution to cope with the challenges of quality inspection in current smart manufacturing, and expect its further applications in the PCB industry.

show abstract

Section: Multimodal Multitask Foundation Modelmentioning

confidence: 99%

Integrated Inspection on PCB Manufacturing in Cyber–Physical–Social Systems

Wang

Cao

et al. 2023

IEEE Trans. Syst. Man Cybern, Syst.

View full text Add to dashboard Cite

show abstract

“…Deep learning (DL), a subarea of ML, allows studying the underlying resources in data from various processing layers using neural networks, similar to the human brain (Goodfellow et al [36]). As of 2010, DL has attracted immense attention in many fields, especially image recognition and speech recognition (Ferreira et al [37], Schmidhuber [38]).…”

Section: Artificial Intelligence and Important Subareasmentioning

confidence: 99%

The Importance of Digital Transformation in International Business

et al. 2022

View full text Add to dashboard Cite

This study was developed under the scope of a Portuguese project focused on the entrepreneur’s perspective and perception on the internationalization process of his company: more specifically, about the factors that enhanced the company entry into foreign markets as well as the constraints found in this process. This work focuses on the importance of using digital transformation to integrate technological tools in international business practice and strategy and the obstacles encountered with introducing these new technologies. This study aims to determine the relationships between technology categories and obstacles. The final goal is to assess the impact of these characteristics of the companies by the sector of economic activity, size, and percentage of profits resulting from international expansion. A questionnaire was designed and sent by email to 8183 companies from the AICEP database, distributed by three main activity sectors. A total of 310 valid answers were gathered from the Portuguese internationalized companies. The research limitations are related to the reduced number of interviews. These interviews showed that managers were not aware of the concept of digital transformation and misunderstood the use of digital technologies in the internationalization process of the business. This limitation can add some bias to the qualitative results. In addition to these limitations, the number of responses per sector was also not homogeneous. The practical implications of this study are that managers and top-level executives can use that to better understand how companies could use digital tools and what obstacles they should avoid when they want to internationalize their business. This paper is one of the first research contributions to analyze the impact of digital transformation in the internalization of Portuguese companies.

show abstract

“…In recent years, due to the rapid development of RNN in natural language processing (NLP), many researchers attempt to apply RNN in SLR [26], [27], [44]. To model the underlying sign language more precisely, different types of encoder-decoder networks, such as the Transformer model, are also employed [45]- [49]. In addition, many new ideas are introduced in SLR.…”

Section: B the Sequential Modulementioning

confidence: 99%

SignBERT: A BERT-Based Deep Learning Framework for Continuous Sign Language Recognition

2021

View full text Add to dashboard Cite

Continuous sign language recognition (CSLR) is a very challenging task in intelligent systems, since it requires to produce real-time responses while performing computationally intensive video analytics and language modeling. Previous studies mainly focus on adopting hidden Markov models or recurrent neural networks with a limited capability to model specific sign languages, and the accuracy can drop significantly when recognizing the signs performed by different signers with non-standard gestures or non-uniform speeds. In this work, we develop a deep learning framework named SignBERT, integrating the bidirectional encoder representations from transformers (BERT) with the residual neural network (ResNet), to model the underlying sign languages and extract spatial features for CSLR. We further propose a multimodal version of SignBERT, which combines the input of hand images with an intelligent feature alignment, to minimize the distance between the probability distributions of the recognition results generated by the BERT model and the hand images. Experimental results indicate that when compared to the performance of alternative approaches for CSLR, our method has better accuracy with significantly lower word error rate on three challenging continuous sign language datasets.INDEX TERMS bidirectional encoder representations from transformers, continuous sign language recognition, deep learning, video analytics I. INTRODUCTION

show abstract

DeSIRe: Deep Signer-Invariant Representations for Sign Language Recognition

Cited by 10 publications

References 36 publications

Integrated Inspection on PCB Manufacturing in Cyber–Physical–Social Systems

Integrated Inspection on PCB Manufacturing in Cyber–Physical–Social Systems

The Importance of Digital Transformation in International Business

SignBERT: A BERT-Based Deep Learning Framework for Continuous Sign Language Recognition

Contact Info

Product

Resources

About