(Press-News.org) Affective computing, proposed by Picard in 1997, aims to endow computational systems with the ability to recognize, interpret, and respond to human emotions. Early studies relied primarily on behavioral cues such as facial expressions and voice tone for modelling affective states.
Affective computing has entered a new phase—wearable devices are capable of continuously acquiring multimodal physiological signals from multiple sensor channels that differ in terms of sampling frequency, physiological origin, and signal characteristics.
In a new review published in Intelligent Sports and Health, a duo of researchers from the Department of Psychological and Cognitive Sciences at Tsinghua University reported the current landscape in data processing flow, multimodal fusion strategy, and model architecture. They also summarized the development status and challenges of deep learning methods in the field of affective computing.
“Both public datasets and self-collected data play important roles in affective computing studies,” shares co-author Dan Zhang. “They show high consistency in terms of modality, device, signal length, number of subjects, and label acquisition. For example, they both make extensive use of common physiological modalities such as EDA and HR and rely mainly on commercially available wearable devices (e.g., Empatica E4) for their measurements.”
Notably, most of the labels are acquired by self-assessment tools (e.g., SAM scales), and the number of subjects is generally on the scale of tens of people.
“Some of the self-collected data incorporate sports-related scenarios, such as walking simulations, which record multimodal signals (including EDA, ACC, HR, etc.) to capture affective changes in individuals during physical activity,” adds Zhang. “These types of data hold potential value for applications in the sports field, such as monitoring emotional fatigue during training or assessing athletes’ emotional regulation capabilities under competitive pressure.”
Further, multimodal fusion in affective computing can be implemented at different stages of the modelling pipeline, including feature-level, model-level, and decision-level fusion. “Feature-level fusion provides a simple and easy-to-operate mechanism, model-level fusion can capture cross-modal interactions within the network structure, and decision-level fusion allows for independent processing of each modality,” syas co-author Fang Li. “The selection of fusion strategies typically depends on the temporal characteristics, complementarity, and reliability of the involved modalities, as well as the complexity of the classification task.”
The authors noted that deep learning methods can extract and model feature representations of input data. Such models include CNNs for extracting local features, LSTM for capturing temporal dependencies, and transformers for addressing temporal dependencies over long distances through a self-attention mechanism.
###
Contact the author: Fang Li, Dan Zhang, Department of Psychological and Cognitive Sciences, Tsinghua University, Beijing, China. dzhang@tsinghua.edu.cn.
The publisher KeAi was established by Elsevier and China Science Publishing & Media Ltd to unfold quality research globally. In 2013, our focus shifted to open access publishing. We now proudly publish more than 200 world-class, open access, English language journals, spanning all scientific disciplines. Many of these are titles we publish in partnership with prestigious societies and academic institutions, such as the National Natural Science Foundation of China (NSFC).
END
A newly discovered, carnivorous lizard ostensibly represents what most casual onlookers would perhaps perceive to be a dinosaur; however, it is in fact a precursor of the modern crocodile.
Named Tainrakuasuchus bellator, partially derived from Latin meaning for “warrior” or “fighter”, this armour-plated reptile lived 240 million years ago just before the dinosaurs.
And, as a Pseudosuchia (the precursors of modern crocodiles and alligators), it was among a group of the top, most dominant predators of its time in the Triassic Period.
This particular new species – detailed, today, in the peer-reviewed Journal of Systematic Palaeontology – was approximately ...
As artificial vision systems evolve, bridging the gap between sensing and processing remains a key challenge. Now, researchers from the University of Science and Technology of China, led by Prof. Yong Yan and Prof. Haiding Sun, have developed a reconfigurable bioinspired vision sensor using GaN/AlN quantum-disk-in-nanowires (QD-NWs) that emulates the human retina’s dual-cell system—delivering in-sensor computing for high-accuracy human action recognition (HAR).
Why This Bioinspired Sensor Matters
Dual-Mode ...
Nuclear fusion, which operates on the same principle that powers the Sun, is expected to become a sustainable energy source for the future. To achieve fusion power generation, it is essential to confine plasma at temperatures exceeding one hundred million degrees using a magnetic field and to maintain this high-energy state stably. A key factor in accomplishing this is the electric potential inside the plasma. This potential governs the transport of particles and energy within the plasma and plays a crucial role in establishing a state in which energy is effectively confined and prevented ...
CHAMPAIGN, Ill. — A naturally occurring protein that tends to be expressed at higher levels in breast cancer cells boosts the effectiveness of some anticancer agents, including doxorubicin, one of the most widely used chemotherapies, and a preclinical drug known as ErSO, researchers report. The protein, FGD3, contributes to the rupture of cancer cells disrupted by these drugs, boosting their effectiveness and enhancing anticancer immunotherapies.
The discovery is described in the Journal of Experimental & Clinical Cancer Research.
The new findings were the happy result of experiments involving ErSO, an experimental drug that killed 95-100% of estrogen-receptor-positive ...
A widely-used, inexpensive gout drug could reduce heart attacks and strokes in people with cardiovascular disease, according to a new Cochrane review.
The review examined the effects of low doses of colchicine, a drug used to treat gout, and found no increase in serious side effects.
Cardiovascular disease is often driven by chronic low-grade inflammation, which contributes to recurrent cardiovascular events such as heart attacks and strokes. Colchicine has anti-inflammatory properties that make it a promising option for people with heart disease.
A promising effect on cardiovascular risk
The review included 12 randomized controlled trials involving nearly 23,000 ...
EMBARGOED: Hold for release until Wednesday, Nov. 12 at 18:30 ET / 23:30 UK time
Headache disorders affect 3 billion people worldwide—nearly one in every three people, ranking sixth for health loss in 2023
Women experienced more than twice the headache-related health loss of men and spent a greater share of their lives with headache symptoms.
Migraine caused nearly 90% of all disability linked to headache disorders, even though tension-type headache was more common.
More than one-fifth of the global headache burden ...
ROCHESTER, Minn. — Mayo Clinic researchers have developed a new tool that can estimate a person's risk of developing memory and thinking problems associated with Alzheimer's disease years before symptoms appear. The research, published in The Lancet Neurology, builds on decades of data from the Mayo Clinic Study of Aging — one of the world's most comprehensive population-based studies of brain health.
The study found that women have a higher lifetime risk than men of developing dementia and mild cognitive impairment (MCI), a transitional stage between healthy aging and dementia ...
Extending anti-clotting drugs beyond the initial treatment period of at least 90 days after a first blood clot is linked to lower rates of new clots developing compared with stopping treatment, finds a US study published by The BMJ today.
Extended treatment was also linked to higher rates of major bleeding (a common side effect of anti-clotting drugs) but the authors say continued use results in an overall “net clinical benefit.”
Guidelines recommend anti-clotting treatment for at least 3-6 months for patients with venous thromboembolism ...
Exposing children to e-cigarettes compromises their human rights. These products should be regulated in a way that puts children’s best interests first and protects them from the harms associated with nicotine consumption in all its forms, argue experts in The BMJ today.
Children are now using e-cigarettes at higher rates than adults, write Tom Gatehouse and colleagues. Globally, the World Health Organization (WHO) estimates that 7.2% of children aged 13-15 currently use e-cigarettes. Data indicate use in this group is nine times that of adults, in countries which monitor both groups.
As ...
The rate of high blood pressure (hypertension) in children and adolescents under 19 has nearly doubled, increasing from about 3% in 2000 to over 6% in 2020.
More than 9% of children and adolescents have masked hypertension — high blood pressure that only shows up with out-of-office tests, meaning hypertension could be undetected during regular checkups.
Nearly 19% of children and adolescents with obesity have hypertension, eight times higher than the prevalence of hypertension in those considered a healthy weight.
Approximately 8% of children and adolescents now have prehypertension, a warning sign of potential progression to hypertension, which ...
