UTILIZING MACHINE LEARNING AND IMAGE PROCESSING TO DETECT SIGNS OF STRESS IN INDIVIDUALS

— This study's main objective is to identify stress in the human body using vivid machine learning and image processing techniques. Our system is an improved version of earlier stress detection systems that lacked personal counseling or live detection. Instead, it detects employees' levels of both physical and mental stress and offers appropriate stress management strategies through a survey form. Additionally, our system includes periodic analysis of employees and live detection. To make the most of employees during working hours, our approach is mainly concerned with stress management and fostering a positive, flexible work environment.


INTRODUCTION
Stress management systems are necessary for detecting stress levels that affect our socio-economic situation.According to the World Health Organization, stress is a mental health disorder that affects one out of every four people (WHO).Mental and financial troubles, as well as a lack of clarity at work, bad working relationships, despair, and, in extreme situations, death, are all symptoms of human stress.This necessitates the provision of therapy to help stressed people manage their stress.While it is impossible to totally eliminate stress, taking preventative measures may help you cope.Only medical and physiological peoples can now determine whether or not someone is depressor (stressed).A questionnaire is one of the most used methods for detecting stress.This technique relies primari-ly on individual responses; people will be hesitant to communicate whether or not they are worried.Automatically de-tecting stress lowers the likelihood of health problems and improves society's wellbeing.This involves the creation of a scientific approach for assessing stress levels in people using physiological markers.Since stress is such a signifi-cant societal contribution, a variety of approaches for detecting it have been investigated.It enhances people's quality of life, according to Ghaderi Tal.Stress was assessed using data from respira-tion, heart rate (HR), face electromyog-raphy (EMG), Galvanic skin response (GSR) foot, and GSR hand, with the finding that parameters related to the respiratory process are critical in stress detection.Maria Viqueira et al. present a method for anticipating mental stress that relies only on GSR as a physiologi-cal sensor and uses a standalone stress detecting device.Electrocardiograms alone were utilized by David Liu and colleagues to predict stress levels (ECG).The effectiveness of multimodal sensors in detecting stress in working individuals is investigated experimental-ly.Sensor data from pressure distribu-tion, heart rate, blood volume pulse (BVP), and electro dermal activity is used in this investigation (EDA).In ad-dition, an eye tracker sensor is used, which analyses eye movements in con-nection with stressors such as the Stroop word test and information regarding pick-up tasks.
Nowadays, the IT industry is creating a new standard in the market by introducing new technology and goods.Employee stress levels were also found to set the bar high in this research.Despite the fact that many companies provide mental health benefits to their workers, the problem remains out of control.In this study, we attempt to go further into the topic by attempting to identify stress patterns in working employees in businesses.We plan to use image processing and machine learning methods to analyse stress patterns and narrow down the elements that greatly influence stress levels.Machine learning approaches such as KNN classifiers are used to categorise stress.The employee's picture is snapped by the camera, which acts as input, and image processing is employed at the first step for detection.Picture processing is used to improve an image or extract relevant information from it by converting the image to digital form and executing operations on it.By taking a picture from video frames as input and producing an image or attributes related with that image as output.
Image processing consists of three phases: • Importing the image using photo capture software • Analysing and altering the picture • As a result of the output, a picture that has been modified or a report based on image analysis.
Machine learning, which is an application of artificial intelligence, gives the system the capacity to autonomously learn and develop from self-experiences without being explicitly designed (Artificial Intelligence).Machine Learning creates computer programs that can access data and utilize it to learn for themselves.Explicit programming produces a mathematical model based on "training data" to accomplish the work based on predictions or judgments using Machine Learning.Image mining is a technique for extracting hidden data, associating picture data, and discovering new pat-terns not visible in the image.This category includes image processing, data mining, machine learning, and datasets.According to conservative estimates in medical literature, stress is the cause of 50-80 percent of all physical diseases.Stress is regarded to be the major cause of cardiovascular disease.Stress is associated to diabetes, ulcers, asthma, migraine headaches, skin issues, epilepsy, and sexual dysfunction.Each of these conditions, as well as a slew of others, has a psychosomatic basis (i.e., induced or exacerbated by mental factors like stress).There are three impacts of stress: • Guilt, humiliation, anxiety, anger, and frustration are some of the subjective symptoms of stress.Individuals may also experience fatigue, tenseness, nervousness, irritability, moodiness, or loneliness.calm, and stressed/anxious).In order to measure emotion expression more objectively, the study focused mostly on non-voluntary and semi-voluntary facial signals.Eye-related events, mouth activity, head motion characteristics, and heart rate assessed via camera-based photo-plethysmography were also investigated.In each experimental phase, a feature selection technique was used to pick the most robust characteristics, followed by classification algorithms that discriminated between stress/anxiety and neutral states with reference to a relaxed condition.In addition, a ranking transformation based on self-reports was presented to study the relationship between face attributes and a participant's reported stress/anxiety level.Specific facial signals generated from eye, mouth, head, and camera-based cardiac activity acquire excellent accuracy and are acceptable as discriminative markers of stress and anxiety, according to the findings.

Detection of Stress Using Image Processing and Machine Learning Techniques AUTHORS: Nisha Raichur, Nidhi Lonakadi, Priyanka Mural
Stress is an uncomfortable state of emotional arousal that individuals feel in settings such as sitting in front of a computer for lengthy periods of time.Computers have become a way of life; we spend so much of our time on them that we are more impacted by the ups and downs they create.One cannot totally avoid using computers for work, but one should at least limit his or her use if he or she is concerned about being stressed at a certain moment.Monitoring a person's mental state while working in front of a computer for an extended period of time is critical for their safety.This research uses real-time nonintrusive movies to assess a person's emotional state by analyzing their facial expression.Each video frame contains a distinct feeling, and the stress level is determined in the hours after the video recording.We use a method that enables us to train a model and compare differences in feature prediction.Theano is a Python framework aimed at speeding up the execution and development of the linear regression model, which is employed as a deep learning technique in this case.The results of the experiments reveal that the devised method works effectively with a generic model of all ages.

Techniques for Predicting Stress in
Working Employees Using Machine Learning AUTHORS: U. S. Reddy, A. V. Thota and A. Dharun Stress problems are a widespread problem among today's working IT professionals.Employees are more likely to experience stress when their lifestyles and work environments change.Despite the fact that many sectors and corporations provide mental health-related programs and attempt to improve the office environment, the problem remains out of Control.In this research, we will use machine learning approaches to examine stress patterns in working people and to identify the elements that have a significant impact on stress levels.Data from the OSMI mental health survey 2017 answers of working professionals in the IT sector were used to help with this.After proper data cleaning and pre-processing, we used a variety of Machine Learning approaches to train our model.The accuracy of the models mentioned above was determined and compared.Among the models used, boosting had the best accuracy.Gender, family history, and the availability of health benefits in the job were found as key characteristics that impact stress using Decision Trees.With these findings, businesses may focus their efforts on reducing stress and providing a more pleasant working environment for their workers.The diagnosis of chronic stress is crucial in predicting and lowering the risk of cardiovascular disease.This project is pilot research with the goal of establishing a technique for identifying short-term psychophysiological alterations using HRV properties.The goal of this pilot project is to identify and acquire insight into a collection of characteristics that might be utilized to detect psychophysiological alterations associated with chronic stress.Images, noises, mental activities, and rest were used to evoke four distinct forms of arousal, which were then identified using linear and non-linear HRV characteristics from electrocardiograms (ECG) obtained by the wireless wearable ePatch® recorder.Sample entropy, detrended fluctuation analysis, and normalized high frequency features were used to get the greatest identification rates for the neutral stage (90 percent), acute stress stage (80 percent), and baseline stage (80 percent).It was discovered that standardizing nonlinear HRV variables for each participant was a crucial component in improving classification outcomes.
5. Healthy Office: Using smartphones and wearable sensors, employees may recognize their moods at work.AUTHORS: Zenonos, A., Khan, A., Kalogridis, G., Vatsikas, S., Lewis, T., Sooriyabandara Workplace stress, anxiety, and depression are detrimental to employees' health and productivity, and they are costly.Sensor technologies, such as smartphones and wearables with physiological and movement sensors, have been the focus of recent research in this field.In this paper, we look at the feasibility of deploying such gadgets for mood detection in the workplace.Every two hours, we propose a new mood detection framework that can recognize five intensity levels for eight distinct kinds of emotions.We also propose a smartphone app ('Healthy Office') that is meant to promote formal self-reporting and give data for our model.In a smallscale user research, we gather wearable sensing data in an office setting to assess our technology.Our trials have shown encouraging results, enabling us to accurately distinguish different types of emotions.

Problem Statement
Image capturing is automated, so it collects photographs whenever a typical behavior occurs.The detecting of system will be fooled.If the picture is distorted while being captured, the system will provide incorrect findings.Continued picture capture results in massive, useless datasets.Detection will become more time demanding or incorrect as a result of the auto collected picture datasets.

Existing System
Work on stress detection in the present system is based on digital signal processing, which takes into account of galvanic skin reaction, blood volume, pupil dilation, and skin temperature.Other research on this topic relies on a variety of physiological signals and visual aspects (eye closure, head movement) to assess a person's stress levels while they are at work.These measures, on the other hand, are obtrusive and uncomfortable in practice.Every sensor reading is compared to a stress index, which is a number that is used to determine the amount of stress.

Advantages of Proposed System
• The output of image analysis is a transformed image or a report.• By providing proactive stress management solutions, the Stress Detection System assists workers in dealing with challenges that cause stress.• We'll take pictures of staff at regular intervals, and then give them the traditional survey forms.

Module Description
The following modules are used in this project: Neighbour Classification) • Here, we will use haarcas-

cade_frontalface_default.xml (classifier)
• User: The first may be registered by the user.For future contact, he needed a valid user email and cell phone number while registration.After the user has registered, the customer may be activated by the administrator.After the customer has been activated by the admin, the user may log into our system.The user must first provide the system with an image as input.The image's attributes and relevant emotion will be extracted using the Python package.It is also feasible to recognize many faces in a same picture.The stress level will be shown by facial expressions such as sadness, anger, and so on.After the picture processing is complete, we will begin the live feed.We can also see the face expressions of many people on the live broadcast.TensorFlow live stream is more accurate and faster than tensor flow live stream.After that, we'll import the dataset in order to calculate the KNN classification accuracy precession scores.
• Admin: With his credentials, the administrator can log in.He may activate the users after logging in.Only the activated user may access our apps.The admin may dynamically set the project's training and testing data in the code.In a hidden frame, the admin may see all of the users' identified results.He can discern the emotions of the photos by clicking a hyperlink on the screen.The results of the Knearest neighbor classification discovered may also be seen by the admin.The spreadsheet with the data.We may extend the dataset size based on fictitious values by allowing authorized people to do so.
• Data Pre-process: Dataset contains a grid view of an existing dataset with numerous properties; however, Property Extraction produces a newly designed dataset with only numerical input variables as a result of Principal Component Analysis feature selection, which transforms to six principal components: Condition (No stress, Time pressure, Interruption), Stress, Physical Demand, Performance, and Frustration.
• Machine Learning: K-Nearest Neighbour (KNN) is a classification and regression analysis algorithm.It's a supervised learning system for determining whether or not a person need therapy.The dependent variable is classified by KNN based on how similar it is to a comparable instance from previously collected data.KNN A statistical model with a binary dependent variable is known as classification.KNN is used to estimate the parameters of a KNN model in classification analysis.A binary KNN model, mathematically, contains a dependent variable with two potential values, which is represented by an indicator variable, with the two values labelled "0" and "1".

Project Design and Analysis
3.1 Architecture • A bubble chart is another name for a DFD.It is a basic graphical formalism that may be used to depict a system in terms of the data it receives, the processing it does on that data, and the data it generates as output.• The data flow diagram is one of the most important modelling tools (DFD).It's used to symbolize the numerous parts of the system.These components include the system process, the data used by the process, an external entity that interacts with the system, and the information flows in the system.• DFD depicts how data flows through the system and is transformed by a sequence of transformations.It's a graphical representation of data flow and the changes that occur when data goes from input to output.• DFD is sometimes referred to as a bubble chart.At any level of abstraction, a DFD may be used to depict a system.DFD may be divided into levels, each representing a different degree of information flow and functional detail. Algorithm: 1) Start 2) Using a camera, capture the input image/object.3) If the image/object is clear, it may be used as input, or it can be told to be rechecked.4) After the input picture has been recognized, it displays the object/stress images in the form of emotions such as joyful, sad, furious, and so on.5) Stop

Input Design
The information system and the users are linked via the input design.It entails defining data preparation criteria and procedures for turning transaction data into a format that may be used.Processing may be done in one of two ways: by scanning a computer for data from a written or printed document, or by having individuals manually enter the data into the system.The goals of input design are to decrease the quantity of data needed, control mistakes, reduce delays, remove superfluous steps, and simplify the process.The input is meant to give security and convenience while maintaining confidentiality.The following aspects were considered by Input Design: • What data should be given as input?• How should the data be organized or coded?• A conversation to assist operational people in offering feedback.• How to prepare input validations and what to do if a mistake occurs.

Objectives of input Design
• The process of translating a user-oriented description of an input into a computer-based system is known as input design.This design is critical for avoiding data entry mistakes and directing management in the right way for collecting accurate information from the computerized system.
• It is accomplished by designing user-friendly data input panels that can manage enormous amounts of data.The purpose of input design is to make data entering simpler and error-free.
The data entering panel is set up in such a manner that you may execute all of the data manipulations.It also allows you to see your records.• It will validate the data after it has been input.Screens may be used as v input information.Appropriate messages are sent when required, ensuring that the user is never caught off guard.As a result, the goal of input design is to produce an easy-tofollow input layout.

Output Design
A quality output is one that satisfies the end user's needs and shows information clearly.Any system's processing results are conveyed to users and other systems through outputs.In output design, it's decided how information will be displaced for immediate use, as well as the format.hard copy output.It is the most important and direct source information to a user.The system's relationship with the user is improved via efficient and intelligent output design.
• Computer output should be built in an orderly, well-thought-out manner; the proper output must be generated while ensuring that each output part is designed in such a way that users will find the system easy to use.They should identify the exact output that is required to satisfy the criteria while analysing and designing computer output.• Decide on how you'll deliver the data.• Create a paper, report or other document that includes the information from the system.
An information system's output form should achieve one or more of the following goals.
• Disseminate information regarding previous actions, present status, or future estimates.• Alert people to significant events, opportunities, challenges, or warnings.• Initiate a reaction.
• Verify a decision.

CONCLUSION
The Tension Detection System forecasts worker tension by using authorized users' collected images, ensuring the system's safety.Upon authenticating, the user automatically takes a picture based on a predetermined time frame.Utilizing standard convergence and image processing techniques, the obtained images are employed to assess the user's level of stress.After that, the system will analyze the stress levels using machine learning techniques, producing more effective results.

Future Enhancement
Biomedical wearable sensors and Internet of Things technology have shown to be a successful combination in the healthcare sector.The benefits of using these devices have benefited patients as well as medical professionals.A few advantages include early identification of health problems, prompt medical assistance via remote monitoring and telecommunication, an emergency alert system to notify the personal physician and caregiver, and more.Better health aids are guaranteed by the proposed work on building a multimodal IoT system, which constantly monitors and provides timely feedback on stress levels.It would be fascinating to build on this work in the future by include more physiological characteristics, like an activity recognition system, and by using machine learning techniques to the development of a stress detection model.