DevRelax is a desktop application designed and developed using machine learning technologies to detect and assist in relieving users stress levels and alleviating emotions.
Stress has become an all-encompassing concern in today's fast-paced world, casting a shadow over individuals' health, well-being, and productivity. The extensive ramifications of stress, including burnout, heart disease, and even premature death, underscore the urgency of addressing this issue. Paniker et al.'s survey [1] explores the application of machine learning techniques to detect stress, representing a key avenue in this endeavor. This review paper synthesizes insights from four core components related to stress and emotion detection and management, offering a comprehensive overview of the landscape. Notably, the shift towards remote work, especially in the IT industry, introduces new stressors, such as isolation, exacerbating stress levels and associated health problems [2].
Stress, as a multifaceted phenomenon, can lead to a wide array of detrimental consequences, from mental health issues to severe depression and, in the worst cases, suicide. Traditional methods for stress detection often rely on subjective symptom-based assessments that necessitate expert intervention. Nevertheless, recent technological advances, notably in the realm of machine learning, hold promise for enhancing stress detection and management. Gedam et al.'s review paper [3] showcases how stress can be effectively and accurately detected through wearable sensors and machine learning techniques. Globally, researchers have explored multiple approaches to stress detection, encompassing facial expression recognition [4], speech analysis [5], and wearable devices [3]. While these methods offer valuable insights, they often focus on isolated physiological factors, indicating the need for a more integrated approach, particularly for high-stress environments like the IT industry.
Research gaps exist concerning the accessibility of stress detection tools, such as headgear for brain wave analysis. Not all individuals may have access to such equipment, potentially limiting the reach of stress detection and management solutions.
Some stress management solutions primarily rely on audio waves and music for stress relief, which may limit their effectiveness. A gap exists in providing a more comprehensive approach to managing stress beyond audio-based interventions.
Existing systems often require users to actively engage in stress detection processes. There is a gap in developing background applications that seamlessly detect stress without requiring additional user actions (Running in the Background seamlessly).
Research gaps include the critical and immediate need for near-real-time stress and emotion data collection, significantly enhancing the effectiveness of stress management solutions compared to methods that involve manual user actions or video logs.
Research gaps involve the application of Multi-Objective Reinforcement Learning (MORL) techniques for providing activity recommendations that address both stress relief and emotion alleviation simultaneously. This represents a promising area for improvement.
While some research emphasizes physiological stress markers like heart rate monitoring, gaps exist in integrating context-awareness into stress management solutions, allowing for personalized and timely stress reduction recommendations based on specific user situations and needs.
The primary research problem addressed in this study is the need to detect users' stress levels and emotions without interrupting their daily work while using common equipment. This issue arises due to the prevalence of prolonged computer screen work, particularly among IT professionals, which often leads to heightened stress levels and emotional strain. The challenge is to provide effective stress relief and emotion alleviation solutions in this context. To address this core problem, four sub-research problems have been identified, encompassing stress detection from keyboard dynamics, external mouse with HR sensor, facial feature analysis, and computer-based activity recommendations. These sub-problems collectively aim to provide a comprehensive approach to managing stress and emotional well-being for IT professionals without disrupting their work routines.
The research problem at hand centers on the imperative need to detect users' stress levels and emotions without disrupting their daily work activities, employing readily available equipment. In the fast-paced world of information technology, characterized by extended periods of computer screen work, stress has become a prevalent concern. The proposed solution adopts a multifaceted approach by addressing four key sub-research problems: stress detection through keyboard dynamics, real-time emotion detection through facial feature analysis, and personalized recommendations for stress relief and emotion alleviation. This holistic approach aims to empower IT professionals by seamlessly and unobtrusively monitoring their well-being during prolonged computer screen work, offering context-specific interventions to enhance their overall quality of life without work disruption.
Develop a robust stress detection system that accurately assesses users' stress levels.
Implement an emotion detection component that can precisely identify and evaluate users' emotional states.
Create a user-friendly desktop application with an intuitive interface, ensuring a seamless user experience.
Enable real-time monitoring of stress levels and emotions, allowing for timely interventions.
Provide personalized recommendations and interventions to assist users in managing their stress and emotions effectively.
Ensure that the system operates discreetly in the background, allowing users to perform their tasks without disruption while benefiting from stress and emotion management.
The system architecture at the core of this proposed
solution aims to develop a machine learning and
reinforcement learning-based application tailored for IT
workers, aggregating user data from keyboard dynamics,
heart rate dynamics, and facial dynamics to provide
individualized stress-relief activities. To ensure
precise stress detection and individualized stress
reduction recommendations, performance is assessed
through user testing and cross-validation.
The proposed solution comprises four main
components: detecting stress through keystroke dynamics,
detecting stress via an external mouse using an HRV
sensor, detecting stress facial analysis using a webcam,
and providing personalized stress-reduction
recommendations. These components work together to offer
a more efficient method of stress management tailored to
specific needs and preferences. The system flow includes
users performing their regular work tasks, with the
components monitoring keyboard and mouse data,
processing it through machine learning models, and
generating predictions for the user's stress level. The
recommendation system then evaluates the data and
suggests stress-relieving activities for the user to
help reduce their stress levels.
During this phase, we carefully assessed and selected a research topic that held significant relevance to our project's goals, ensuring that our research direction was well-founded.
We created a project charter and proposal, outlining the scope, objectives, and a detailed plan for our research project, serving as a roadmap for the entire research journey.
During this phase, we presented an initial overview of our research progress to stakeholders, providing a glimpse of our project's development and key findings.
This document summarized the current status of our project, offering a snapshot of where we stood in terms of research goals and accomplishments.
We compiled a comprehensive research paper, encompassing our research methodology, findings, and conclusions, thereby showcasing the core outcomes of our research.
During this phase, we updated the project's status in our second status document, highlighting any changes in the project's status and findings that had occurred since the initial status document.
During this phase, we prepared final reports that consolidated all aspects of our research, including detailed results and recommendations, providing a conclusive overview of our research efforts.
Our second progress presentation presented comprehensive results and findings to the panel, offering a deep dive into the outcomes of our research.
During this phase, we critically assessed the overall success and impact of our research project, identifying its strengths and areas for potential improvement, which will inform future research directions.
We reviewed and evaluated the logbook or research diary, ensuring that our research activities and insights were properly documented for future reference and transparency.
During this phase, we evaluated the project website's effectiveness in disseminating information about our research project and promoting it to a wider audience, assessing its role in our outreach efforts.
