1. A wearable electroencephalogram-based emotion recognition system (100), the system (100) comprising: a wearable headset (102) positioned on a user's head for emotional recognition; a plurality of electrodes (104) disposed within the wearable headset (102) and configured to collect electroencephalogram signals from a user’s scalp; a processing unit (106) embedded within the wearable headset (102) and configured to process real-time electroencephalogram signals for emotion recognition, wherein the processing unit (106) further comprises: a signal input module (108) configured to receive real-time electroencephalogram signals from the plurality of electrodes (104); a pre-processing module (110) configured to clean, remove noise and filter unwanted signal from the received signals to enhance signal quality; a signal quality checking module (112) configured to check the quality of the signals; a feature extraction module (114) configured to extract the signal related features from the pre-processed signals; a brain mapping module (116) configured to generate a representation of brain regions based on the extracted features; a brain analysing module (118) configured to analyse the generated brain representation to identify spatial and spectral patterns in brain regions; a detection module (120) configured to detect the presence of an emotion based on the identified patterns; a classification module (122) configured to classify the detected emotion into multiple predefined emotion categories; and an output module (126) configured to transmit extracted features, mapping, detected emotion and classified emotional state.

2. The system (100) as claimed in claim 1, wherein the processing unit (106) further comprises a micro-calibration module (124) configured to continuously adjust electrode drift and inter-session variations in the EEG signals and to provide calibrated signals for improved emotion recognition accuracy.

3. The system (100) as claimed in claim 1, wherein the system (100) operates independent of network connectivity and remote computational resources during real-time emotion recognition.

4. The system (100) as claimed in claim 1, wherein the feature extraction module (114) configured to extract the signal related features including temporal, spectral, spatial, and entropy-based features.

5. The system as claimed in claim 1, wherein the brain analysing module (118) employs dual-attention graph convolution model to analyse the generated brain representations to identify spatial and spectral patterns.

6. The system as claimed in claim 1, wherein the classification module (122) classifies the detected emotion into multiple predefined emotion categories including but not limited to happy, sad, angry, neutral, fear, surprise, disgust and many more.

7. A method (200) for wearable electroencephalogram-based emotion recognition system (100), the method (200) comprising: positioning a wearable headset (102) on a user's head for emotion recognition; collecting real-time electroencephalogram signals from a user’s scalp via a plurality of electrodes (104); processing the real-time electroencephalogram signal for emotion recognition via a processing unit (106) comprising several modules; receiving electroencephalogram signals from the plurality of electrodes (104) via a signal input module (108); cleaning, removing noise and filtering unwanted signal from the received signals to enhance signal quality via a pre-processing module (110); checking the quality of the signals via a signal quality checking module (112); extracting the signal related features from the pre-processed signals via a feature extraction module (114); generating a representation of brain regions using positions of the plurality of electrodes (104) via a brain mapping module (116); analysing the generated brain representations to identify spatial and spectral patterns in brain regions via a brain analysing module (118); detecting the presence of emotion based on identified patterns via a detection module (120); classifying the detected emotion into multiple predefined emotion categories via a classification module (122); adjusting for electrode drift and inter-user variations in the EEG signals and to provide calibrated signals for improved emotion recognition accuracy via a micro-calibration module (124); and transmitting extracted features, mapping, detected emotion signal and classified emotional state via an output module (126).

Description:FIELD OF DISCLOSURE
[0001] The present disclosure generally relates to biomedical engineering and neuro-signal processing technologies, more specifically, relates to a wearable electroencephalogram-based emotion recognition system.
BACKGROUND OF THE DISCLOSURE
[0002] Emotion recognition from electroencephalogram (EEG) signals remains a persistent and multifaceted challenge, particularly in subject-independent scenarios where inherent inter-person variability in brain signal patterns, coupled with session-to-session drift due to physiological changes, fatigue, and environmental factors, leads to drastic performance degradation. Some approaches, including convolutional neural networks (CNNs) and graph convolutional networks (GCNs) achieve impressive accuracies in controlled, subject-specific training and testing environments but catastrophically fail to generalize when deployed across unseen users and even the same user in subsequent sessions, severely limiting their viability for practical, everyday wearable systems. This subject-dependency stems from the highly individualized nature of EEG signatures, which vary not only between individuals but also within the same person over time due to factors like mood fluctuations, cognitive states, and minor head movements. Compounding these issues is the complete lack of effective mechanisms for continual in-situ recalibration, leaving systems vulnerable to electrode drift where contact impedance rises from sweat, motion, skin changes and physiological variability including heart rate artifacts and muscle noise, resulting in unreliable signal fidelity over extended wear periods.
[0003] Conventional approaches based on multi-channel EEG with convolutional neural networks for multi-label classification using time-frequency features, multimodal systems combining EEG signals with additional physiological and visual inputs, and methods utilizing graph-based neural networks to model inter-channel correlations, suffer from several critical limitations in addressing real-world EEG challenges. Many such systems rely on supervised learning frameworks that operate in static, subject-dependent settings with limited preprocessing capabilities, lacking mechanisms for domain adaptation to accommodate variations across different users and sessions, thereby resulting in poor generalization. Additionally, multimodal approaches often require multiple sensing modalities and involve complex and manual artifact handling processes, making them unsuitable for standalone wearable deployment and real-time applications. Graph-based models, although capable of capturing spatial relationships are often implemented as centralized and static solutions requiring high computational resources without provisions for continual calibration, interpretability, and privacy-preserving edge execution. Across these existing technologies, common shortcomings include dependence on subject-specific training without alignment for inter-user variability, absence of autonomous in-situ recalibration to address electrode drift and signal degradation, lack of real-time signal-quality evaluation for dynamic channel selection and artifact control, reliance on offline and cloud-based computation that limits mobility and increases latency, insufficient interpretability of spatial and spectral contributions, and potential privacy risks due to centralized data processing without local data confinement. Existing systems uniformly exhibit subject-dependent performance, training on specific users and faltering across diverse individuals and repeated sessions due to absent domain adaptation mechanisms that could align feature distributions. They lack autonomous in-situ recalibration protocols to combat electrode drift from motion, physiological shifts like fatigue, and signal-quality degradation over time. No integrated quality-gating exists for real-time EEG integrity checks, dynamic channel selection, adaptive re-referencing, and automatic artifact suppression, allowing noise propagation. Inference is confined to offline PC, cloud, high-end hardware, imposing latency, connectivity reliance, and scalability issues unsuitable for mobile wearables. Interpretability is minimal, absent spatial/spectral saliency maps revealing brain-region and frequency-band roles treating models as opaque classifiers unfit for clinical scrutiny. Privacy is compromised by architectures demanding raw EEG upload for processing and personalization, exposing neural data to external risks without local confinement. Collectively, these limitations relegate the inventions to lab-bound prototypes, unable to deliver robust, real-time, privacy-secure emotion recognition in dynamic, user-agnostic wearable context.
[0004] The present invention provides substantial improvements over through its innovative architecture that delivers robust, real-time performance in diverse real-world conditions, far surpassing the constrained, lab-centric capabilities of existing systems. Unlike conventional inventions hampered by subject-specific training and poor cross-user generalization, it ensures consistent accuracy across varied individuals and sessions without retraining overheads. By eliminating reliance on offline and cloud infrastructure, it achieves low-latency operation ideal for mobile wearables, addressing the connectivity and scalability bottlenecks of existing systems. It inherently counters signal degradation and noise through proactive, autonomous adjustments, maintaining reliability over prolonged use where others falter due to unaddressed drift and artifacts. Enhanced transparency into decision-making processes builds trust for practical applications, overcoming the black-box opacity prevalent in current models. Privacy is fortified by design, confining sensitive data locally and mitigating external exposure risks that plague centralized approaches. Overall, these advancements enable seamless transition from prototypes to deployable technology, outperforming prior inventions in generalization, efficiency, robustness, usability, and security.
[0005] Thus, in light of the above-stated discussion, there exists a need for a wearable electroencephalogram-based emotion recognition system.
SUMMARY OF THE DISCLOSURE
[0006] The following is a summary description of illustrative embodiments of the invention. It is provided as a preface to assist those skilled in the art to more rapidly assimilate the detailed design discussion which ensues and is not intended in any way to limit the scope of the claims which are appended hereto in order to particularly point out the invention.
[0007] According to illustrative embodiments, the present disclosure focuses on a wearable electroencephalogram-based emotion recognition system which overcomes the above-mentioned disadvantages or provide the users with a useful or commercial choice.
[0008] An objective of the present disclosure is to provide a wearable EEG emotion engine that achieves subject-independent emotion recognition with high accuracy across diverse users and sessions.
[0009] Another objective of the present disclosure is to enable real-time, low-latency inference on low-power edge hardware, eliminating the offline and cloud dependencies and processing delays.
[0010] Another objective of the present disclosure is to incorporate continual in-situ micro-calibration for autonomous correction of electrode drift and physiological variability, ensuring sustained performance over extended wear periods where prior arts degrade.
[0011] Another objective of the present disclosure is to ensure reliable emotion detection by continuously monitoring and improving signal clarity, overcoming noise and quality issues that plague traditional EEG systems.
[0012] Another objective of the present disclosure is to deliver interpretable saliency maps highlighting brain-region and frequency-band contributions.
[0013] Another objective of the present disclosure is to ensure privacy-preserving operation, mitigating signal transmission risks in centralized prior architectures.
[0014] Another objective of the present disclosure is to construct domain-adaptive brain-region graphs using electrode geometry for enhanced spatial and spectral patterns extraction, surpassing static graph models in adaptability.
[0015] Yet another objective of the present disclosure is to provide a self-supervised feature alignment that dynamically minimizes distribution shifts between source and target EEG feature domains, enabling seamless adaptation without labelled data from new users.
[0016] In light of the above, in one aspect of the present disclosure, a wearable electroencephalogram-based emotion recognition system is disclosed herein. The system comprises a wearable headset positioned on a user's head for emotional recognition. The system also includes a plurality of electrodes disposed within the wearable headset and configured to collect electroencephalogram signals from a user’s scalp. The system also includes a processing unit embedded with the wearable headset and configured to process real-time electroencephalogram signals for emotion recognition, the processing unit further comprises a signal input module configured to receive real-time electroencephalogram signals from the plurality of electrodes, a pre-processing module configured to clean, remove noise and filter unwanted signal from the received signals to enhance signal quality, a signal quality checking module configured to check the quality of the signals, a feature extraction module configured to extract the signal related features from the pre-processed signals, a brain mapping module configured to generate a representation of brain regions using positions of the plurality of electrodes, a brain analysing module configured to analyse the generated brain representation to identify spatial and spectral patterns in brain regions, a detection module configured to detect the presence of emotion based on identified patterns, a classification module configured to classify the detected emotion into multiple predefined categories, an output module configured to transmit extracted features, mapping, detected emotion signals and classified emotional state.
[0017] In one embodiment, the processing unit further comprises a micro-calibration module configured to continuously adjust for electrode drift and inter-user variations in the EEG signals and to provide calibrated signals for improved emotion recognition accuracy.
[0018] In one embodiment, the system operates independent of network connectivity and remote computational resources during real-time emotion recognition.
[0019] In one embodiment, the feature extraction module configured to extract the signal related feature including temporal, spectral, spatial, and entropy-based features.
[0020] In one embodiment, the brain analysing module employs dual-attention graph convolution model to analyse the generated brain representation to identify spatial and spectral patterns.
[0021] In one embodiment, the classification module classifies the detected emotion into multiple predefined emotion categories including but not limited to happy, sad, angry, neutral, fear, surprise, disgust and many more.
[0022] In light of the above, in another aspect of the present disclosure, a method for wearable electroencephalogram-based emotion recognition system. The method comprises positioning a wearable headset on a user's head for emotion recognition. The method includes collecting real-time electroencephalogram signals from a user’s scalp via a plurality of electrodes. The method also includes processing the real-time electroencephalogram signal for emotion recognition via a processing unit, comprising several modules. The method also includes receiving electroencephalogram signals from the plurality of electrodes via a signal input module. The method also includes cleaning, removing noise and filtering unwanted signal from the received signals to enhance signal quality via a pre-processing module. The method also includes checking the quality of the signals via a signal quality checking module. The method also includes extracting the signal related features from the pre-processed signals via a feature extraction module. The method also includes generating a representation of brain regions using positions of the plurality of electrodes via a brain mapping module. The method also includes analysing the generated brain representation to identify spatial and spectral patterns in brain regions via a brain analysing module. The method also includes detecting the presence of emotion based on identified patterns via a detection module. The method also includes classifying the detected emotion into multiple predefined emotion categories via a classification module. The method also includes adjusting for electrode drift and inter-user variations and to provide calibrated signals for improved emotion recognition accuracy via a micro-calibration module. The method also includes transmitting extracted features, mapping, detected emotion signal and classified emotional state via an output module.
[0023] These and other advantages will be apparent from the present application of the embodiments described herein.
[0024] The preceding is a simplified summary to provide an understanding of some embodiments of the present invention. This summary is neither an extensive nor exhaustive overview of the present invention and its various embodiments. The summary presents selected concepts of the embodiments of the present invention in a simplified form as an introduction to the more detailed description presented below. As will be appreciated, other embodiments of the present invention are possible utilizing, alone or in combination, one or more of the features set forth above or described in detail below.
[0025] These elements, together with the other aspects of the present disclosure and various features are pointed out with particularity in the claims annexed hereto and form a part of the present disclosure. For a better understanding of the present disclosure, its operating advantages, and the specified object attained by its uses, reference should be made to the accompanying drawings and descriptive matter in which there are illustrated exemplary embodiments of the present disclosure.
BRIEF DESCRIPTION OF THE DRAWINGS
[0026] To describe the technical solutions in the embodiments of the present disclosure or in the prior art more clearly, the following briefly describes the accompanying drawings required for describing the embodiments or the prior art. Apparently, the accompanying drawings in the following description merely show some embodiments of the present disclosure, and a person of ordinary skill in the art can derive other implementations from these accompanying drawings without creative efforts. All of the embodiments or the implementations shall fall within the protection scope of the present disclosure.
[0027] The advantages and features of the present disclosure will become better understood with reference to the following detailed description taken in conjunction with the accompanying drawing, in which:
[0028] FIG.1 illustrates a block diagram of a wearable electroencephalogram-based emotion recognition system, in accordance with an embodiment of the present disclosure; and
[0029] FIG. 2 illustrates a flow chart of a method, outlining the sequential steps for wearable electroencephalogram-based emotion recognition system, in accordance with an embodiment of the present disclosure.
[0030] Like reference, numerals refer to like parts throughout the description of several views of the drawing.
[0031] The wearable electroencephalogram-based emotion recognition system is illustrated in the accompanying drawings, which like reference letters indicate corresponding parts in the various figures. It should be noted that the accompanying figure is intended to present illustrations of exemplary embodiments of the present disclosure. This figure is not intended to limit the scope of the present disclosure. It should also be noted that the accompanying figure is not necessarily drawn to scale.
DETAILED DESCRIPTION OF THE DISCLOSURE
[0032] The following is a detailed description of embodiments of the disclosure depicted in the accompanying drawings. The embodiments are in such detail as to communicate the disclosure. However, the amount of detail offered is not intended to limit the anticipated variations of embodiments; on the contrary, the intention is to cover all modifications, equivalents, and alternatives falling within the scope of the present disclosure.
[0033] In the following description, numerous specific details are set forth in order to provide a thorough understanding of the embodiments of the present disclosure. It may be apparent to one skilled in the art that embodiments of the present disclosure may be practiced without some of these specific details.
[0034] Various terms as used herein are shown below. To the extent a term is used, it should be given the broadest definition persons in the pertinent art have given that term as reflected in printed publications and issued patents at the time of filing.
[0035] The terms “a” and “an” herein do not denote a limitation of quantity but rather denote the presence of at least one of the referenced items.
[0036] The terms “having”, “comprising”, “including”, and variations thereof signify the presence of a component.
[0037] Referring now to FIG. 1 and FIG. 2 to describe various exemplary embodiments of the present disclosure. FIG. 1 illustrates a block diagram of a wearable electroencephalogram-based emotion recognition system 100, in accordance with an embodiment of the present disclosure.
[0038] The system 100 may include a wearable headset 102. The system 100 may include a plurality of electrodes 104. The system 100 may include a processing unit 106, which further comprises a signal input module 108, a pre-processing module 110, a signal quality checking module 112, a feature extraction module 114, a brain mapping module 116, a brain analysing module 118, a detection module 120, a classification module 122, a micro-calibration module 124 and an output module 126.
[0039] The wearable headset 102 positioned on a user's head for emotional recognition. The wearable headset 102 is ergonomically structured to securely fit different head sizes and maintain stable contact with the scalp during use, thereby enabling reliable acquisition of brain signals.
[0040] In one embodiment of the present invention, the wearable headset 102 is configured to accommodate the plurality of electrodes 104 and the processing unit 106 for real-time signal acquisition and processing within a compact and wearable form factor.
[0041] The plurality of electrodes 104 disposed within the wearable headset 102 and configured to collect electroencephalogram signals from a user’s scalp. The plurality of electrodes 104 are arranged at predefined positions corresponding to different regions of the scalp to enable acquisition of multi-channel brain signals and maintain consistent electrical contact with the scalp to ensure stable and continuous signal acquisition during use.
[0042] In one embodiment of the present invention, the plurality of electrodes 104 are non-invasive electrodes configured for acquisition of EEG signals from the brain.
[0043] The processing unit 106 embedded with the wearable headset 102 and configured to process real-time electroencephalogram signals for emotion recognition. The processing unit 106 is operatively coupled to the plurality of electrodes 104 to receive EEG signals and perform processing locally thereby ensuring reduced latency, improved data privacy.
[0044] In one embodiment of the present invention, the processing unit 106 is a low-power edge processing unit such as ARM Cortex-M and ARM Cortex-A processor.
[0045] The signal input module 108 configured to receive real-time electroencephalogram signals from the plurality of electrodes 104. The signal input module 108 receives continuous multi-channel EEG signals corresponding to different brain regions.
[0046] In one embodiment of the present invention, the signal input module 108 further conditions the received signals into a suitable format for downstream processing and ensures synchronized handling of multi-channel inputs, thereby facilitating seamless transfer of EEG signals to the pre-processing module 110 for subsequent signal refinement.
[0047] The pre-processing module 110 configured to clean, remove noise and filter unwanted signal from the received signals to enhance signal quality. The pre-processing module 110 operates on the received EEG signals to reduce disturbances and irregularities present in the signals and improve clarity prior to further analysis.
[0048] The signal quality checking module 112 configured to check the quality of the signals. The signal quality checking module 112 evaluates the pre-processed EEG signals to determine reliability and usability for further recognition.
[0049] In one embodiment of the present invention, the signal quality checking module 112 further performs dynamic channel selection by excluding signals affected by disturbances, and instability. The signal quality checking module 112 operates using rule-based evaluation of signal consistency and stability rather than a learning-based model, thereby ensuring that only high-quality EEG signals are forwarded for subsequent feature extraction module 114.
[0050] The feature extraction module 114 configured to extract the signal related features from the pre-processed signals. The feature extraction module 114 processes the refined signals to derive meaningful characteristics representing variations in brain activity.
[0051] In one embodiment of the present invention, the feature extraction module 114 configured to extract the signal related feature including temporal, spectral, spatial, and entropy-based features.
[0052] The brain mapping module 116 configured to generate a representation of brain regions using positions of the plurality of electrodes 104. The brain mapping module 116 is further configured to organize the acquired EEG signals into a structured representation reflecting interconnections between adjacent and functionally related brain regions.
[0053] In one embodiment of the present invention, the brain mapping module 116 generates an adjacency structure representing interconnections between the nodes, thereby enabling transformation of the EEG signals into a graph-structured representation suitable for processing by a domain-adaptive graph convolution model, and the constructed graph representation facilitates subsequent extraction of spatial dependencies and region-wise interactions associated with emotional states.
[0054] The brain analysing module 118 configured to analyse the generated brain representation to identify spatial and spectral patterns in brain regions. The brain analysing module 118 analyses the graph-structured representation of EEG signals to determine relationships and interactions between different brain regions and captures variations across different signal characteristics corresponding to brain activity, thereby enabling identification of patterns associated with emotional states.
[0055] In one embodiment of the present invention, the brain analysing module 118 employs dual-attention graph convolution model to analyse the generated brain representation to identify spatial and spectral patterns.
[0056] In one embodiment of the present invention, spatial patterns represent which region of brain is active and spectral patterns represents frequency characteristics of the signals generated by the brain.
[0057] The detection module 120 configured to detect the presence of emotion based on identified patterns. The detection module 120 interprets the patterns and isolate relevant emotional cues.
[0058] In one embodiment of the present invention, the detection module 120 leverages the generated representations to localize significant brain regions and signal components contributing to emotional responses, thereby enabling precise detection of emotion-specific patterns for subsequent classification.
[0059] The classification module 122 configured to classify the detected emotion into multiple predefined categories. The classification module 122 receives the detected emotion-related signals from the detection module 120 and processes the associated feature representations to assign the signals to corresponding emotional states and distinguishes between different emotional categories based on learned patterns.
[0060] In one embodiment of the present invention, the classification module 122 classifies the detected emotion into multiple predefined emotion categories including but not limited to happy, sad, angry, neutral, fear, surprise, disgust and many more.
[0061] In one embodiment of the present invention, the processing unit 106 further comprises the micro-calibration module 124 configured to continuously adjust for electrode drift and inter-user variations in the EEG signals and to provide calibrated signals for improved emotion recognition accuracy.
[0062] The output module 126 configured to transmit extracted features, mapping, detected emotion signals and classified emotional state. The output module 126 organizes and delivers the processed information in a structured format for user interpretation, thereby enabling real-time feedback of emotional states.
[0063] In one embodiment of the present invention, the output module 126 transmits signal information while preserving privacy by restricting exposure of the underlying EEG signals during transmission.
[0064] In one embodiment of the present invention, the system 100 operates independent of network connectivity and remote computational resources during real-time emotion recognition.
[0065] In one embodiment of the present invention, the system 100 uses a self-supervised domain adaptation configured to align feature distributions across different users and sessions, and the adaptation is achieved using techniques including adversarial learning and distribution alignment, thereby enabling subject-independent emotion recognition and improving generalization performance across varying input conditions.
[0066] In one embodiment of the present invention, the system 100 performs subject-independent emotion recognition by learning generalized feature representations that are invariant to individual user characteristics.
[0067] FIG. 2 illustrates a flowchart of a method 200, outlining the sequential steps for wearable electroencephalogram-based emotion recognition system 100, in accordance with an exemplary embodiment of the present disclosure.
[0068] At step 202, position a wearable headset 102 on a user's head for emotion recognition.
[0069] At step 204, collect real-time electroencephalogram signals from a user’s scalp via a plurality of electrodes 104.
[0070] At step 206, process the real-time electroencephalogram signal for emotion recognition via a processing unit 106, comprising several modules.
[0071] At step 208, receive electroencephalogram signals from the plurality of electrodes 104 via a signal input module 108.
[0072] At step 210, clean, remove noise and filter unwanted signal from the received signals to enhance signal quality via a pre-processing module 110.
[0073] At step 212, check the quality of the signals via a signal quality checking module 112.
[0074] At step 214, extract the signal related features from the pre-processed signals via a feature extraction module 114.
[0075] At step 216, generate a representation of brain regions using positions of the plurality of electrodes 104 via a brain mapping module 116.
[0076] At step 218, analyse the generated brain representation to identify spatial and spectral patterns in brain regions via a brain analysing module 118.
[0077] At step 220, detect the presence of emotion based on identified patterns via a detection module 120.
[0078] At step 222, classify the detected emotion into multiple predefined emotion categories via a classification module 122.
[0079] At step 224, adjust for electrode drift and inter-user variations in the EEG signals and to provide calibrated signals for improved emotion recognition accuracy via a micro-calibration module 124.
[0080] At step 226, transmit extracted features, mapping, detected emotion signal and classified emotional state via an output module 126.
[0081] In the best mode of operation of the present invention, the wearable electroencephalogram-based emotion recognition system 100 comprises a wearable headset 102 configured to be positioned on a user’s head for emotional recognition. The wearable headset 102 houses a plurality of electrodes 104 arranged to acquire real-time electroencephalogram signals from different regions of the brain. The plurality of electrodes 104 are operatively coupled to a processing unit 106 embedded within the wearable headset 102, the processing unit 106 being configured as an on-device low-power edge processor to enable real-time signal processing without reliance on external computational resources and comprises several specialized modules. Initially, the signal input module 108 receives real-time electroencephalogram signals from the plurality of electrodes 104. The received signals may contain noise and inconsistencies hence, it is first processed by the pre-processing module 110, which cleans, removes noise, and filters unwanted disturbances to enhance signal quality. Simultaneously the signal quality checking module 112 evaluates reliability of the signals. The pre-processed EEG signals are then forwarded to the feature extraction module 114 to extract signal-related features including temporal, spectral, spatial, and entropy-based features. These extracted features are provided to the brain mapping module 116 to generate a representation of brain regions based on positions of the plurality of electrodes 104. The generated brain representation is then supplied to the brain analysing module 118, which analyses the generated brain representation using a dual-attention graph convolution model to identify spatial and spectral patterns associated with emotional states. Following identified patterns, the detection module 120 detects emotion-related patterns. Upon detection, the classification module 122 classifies the detected emotion into multiple predefined emotion categories including but not limited to happy, sad, angry, neutral, fear, surprise, disgust and many more. Concurrently, the micro-calibration module 124 continuously adjusts for electrode drift and inter-user variations to maintain accuracy and enable subject-independent operation. The classified emotional state, along with extracted features and generated representations, is then transmitted by the output module 126 to provide real-time feedback while maintaining system-level privacy. The system 100 operates in a continuous loop to enable real-time recognition and dynamic updating of the user’s emotional state during ongoing use.
[0082] While the invention has been described in connection with what is presently considered to be the most practical and various embodiments, it will be understood that the invention is not to be limited to the disclosed embodiments, but on the contrary, is intended to cover various modifications and equivalent arrangements included within the scope of the appended claims.
[0083] A person of ordinary skill in the art may be aware that, in combination with the examples described in the embodiments disclosed in this specification, units and algorithm steps may be implemented by electronic hardware, computer software, or a combination thereof.
[0084] The foregoing descriptions of specific embodiments of the present disclosure have been presented for purposes of illustration and description. They are not intended to be exhaustive or to limit the present disclosure to the precise forms disclosed, and many modifications and variations are possible in light of the above teaching. The embodiments were chosen and described to best explain the principles of the present disclosure and its practical application, and to thereby enable others skilled in the art to best utilize the present disclosure and various embodiments with various modifications as are suited to the particular use contemplated. It is understood that various omissions and substitutions of equivalents are contemplated as circumstances may suggest or render expedient, but such omissions and substitutions are intended to cover the application or implementation without departing from the scope of the present disclosure.
[0085] Disjunctive language such as the phrase “at least one of X, Y, Z,” unless specifically stated otherwise, is otherwise understood with the context as used in general to present that an item, term, etc., may be either X, Y, or Z, or any combination thereof (e.g., X, Y, and/or Z). Thus, such disjunctive language is not generally intended to, and should not, imply that certain embodiments require at least one of X, at least one of Y, or at least one of Z to each be present.
[0086] In a case that no conflict occurs, the embodiments in the present disclosure and the features in the embodiments may be mutually combined. The foregoing descriptions are merely specific implementations of the present disclosure but are not intended to limit the protection scope of the present disclosure. Any variation or replacement readily figured out by a person skilled in the art within the technical scope disclosed in the present disclosure shall fall within the protection scope of the present disclosure. Therefore, the protection scope of the present disclosure shall be subject to the protection scope of the claims.
, Claims:I/We Claim:
1. A wearable electroencephalogram-based emotion recognition system (100), the system (100) comprising:
a wearable headset (102) positioned on a user's head for emotional recognition;
a plurality of electrodes (104) disposed within the wearable headset (102) and configured to collect electroencephalogram signals from a user’s scalp;
a processing unit (106) embedded within the wearable headset (102) and configured to process real-time electroencephalogram signals for emotion recognition, wherein the processing unit (106) further comprises:
a signal input module (108) configured to receive real-time electroencephalogram signals from the plurality of electrodes (104);
a pre-processing module (110) configured to clean, remove noise and filter unwanted signal from the received signals to enhance signal quality;
a signal quality checking module (112) configured to check the quality of the signals;
a feature extraction module (114) configured to extract the signal related features from the pre-processed signals;
a brain mapping module (116) configured to generate a representation of brain regions based on the extracted features;
a brain analysing module (118) configured to analyse the generated brain representation to identify spatial and spectral patterns in brain regions;
a detection module (120) configured to detect the presence of an emotion based on the identified patterns;
a classification module (122) configured to classify the detected emotion into multiple predefined emotion categories; and
an output module (126) configured to transmit extracted features, mapping, detected emotion and classified emotional state.
2. The system (100) as claimed in claim 1, wherein the processing unit (106) further comprises a micro-calibration module (124) configured to continuously adjust electrode drift and inter-session variations in the EEG signals and to provide calibrated signals for improved emotion recognition accuracy.
3. The system (100) as claimed in claim 1, wherein the system (100) operates independent of network connectivity and remote computational resources during real-time emotion recognition.
4. The system (100) as claimed in claim 1, wherein the feature extraction module (114) configured to extract the signal related features including temporal, spectral, spatial, and entropy-based features.
5. The system as claimed in claim 1, wherein the brain analysing module (118) employs dual-attention graph convolution model to analyse the generated brain representations to identify spatial and spectral patterns.
6. The system as claimed in claim 1, wherein the classification module (122) classifies the detected emotion into multiple predefined emotion categories including but not limited to happy, sad, angry, neutral, fear, surprise, disgust and many more.
7. A method (200) for wearable electroencephalogram-based emotion recognition system (100), the method (200) comprising:
positioning a wearable headset (102) on a user's head for emotion recognition;
collecting real-time electroencephalogram signals from a user’s scalp via a plurality of electrodes (104);
processing the real-time electroencephalogram signal for emotion recognition via a processing unit (106) comprising several modules;
receiving electroencephalogram signals from the plurality of electrodes (104) via a signal input module (108);
cleaning, removing noise and filtering unwanted signal from the received signals to enhance signal quality via a pre-processing module (110);
checking the quality of the signals via a signal quality checking module (112);
extracting the signal related features from the pre-processed signals via a feature extraction module (114);
generating a representation of brain regions using positions of the plurality of electrodes (104) via a brain mapping module (116);
analysing the generated brain representations to identify spatial and spectral patterns in brain regions via a brain analysing module (118);
detecting the presence of emotion based on identified patterns via a detection module (120);
classifying the detected emotion into multiple predefined emotion categories via a classification module (122);
adjusting for electrode drift and inter-user variations in the EEG signals and to provide calibrated signals for improved emotion recognition accuracy via a micro-calibration module (124); and
transmitting extracted features, mapping, detected emotion signal and classified emotional state via an output module (126).