Mental Health Status Detection through Handwriting Analysis

Abstract

A person\'s handwriting is thought to reflect his/her mental and physical coordination. Handwriting analysis is used to make inferences about a person\'s mental health. Emotions are evoked subconsciously and describe a person\'s mental condition. They inspire, organise, and direct people\'s perceptions, thoughts, and actions. They may be positive or negative. Depression, anxiety, and stress are all expressions of negative emotions. It is vital to recognise negative emotion in a person who may require counselling or psychiatric treatment. The suggested technique aims to create a system that employs machine learning to forecast a person\'s mental health without the need for human intervention. The major goal of this study is to extract characteristics from handwritten text, which will subsequently be normalised, scaled, and utilised to make mental health judgments.

Introduction

I. INTRODUCTION

Mental health is just as important as physical health, as a result of which it is vital to keep a check on a person's mental health which includes depression, anxiety and tension which is commonly occurring among individuals. Recent studies show that by the end of December 2022 20% of Indians are likely to be diagnosed with various mental illnesses in their lifetime.

There are multiple ways in which a person’s mental health can be determined by looking at the respective symptoms or by getting it diagnosed by a psychologist.

Various environmental settings in which an individual takes part in also plays a role. For instance, students spend most of their time in classrooms. classrooms are emotional settings. Students’ emotional experiences can impact on their ability to learn, their engagement in school, and their career choices all We know that some emotions provide a barrier to students’ classroom engagement and test performance.[1-3] For example, academic anxieties, have wide-ranging effects, affecting strategy use, test performance, and subject choice which leads to decline in their cognitive capacity[15].

Handwritten analysis is used  to determine mental illnesses such as depression, anxiety and tension in individuals by using machine learning. Handwritten analysis uses DASS scale and CNN.

The DASS is a 42-item questionnaire primarily used to measure the three related bleak emotional states of depression, anxiety and stress. [11] CNN stands for Convolution neural network, it is a powerful tool by deep learning, because of its ability to handle large amounts of data. [12]

It’s applications go beyond handwriting analysis it is also used in image data, classification prediction problems and regression prediction problems.  Recognizing negative emotions is quite important and handwriting analysis serves this purpose which in simple words is the analysis of handwriting with attempt to determine someone's mental health status. Handwriting analysis has existed since the nineteenth century.

Handwriting analysis uses various features to analyse these include baseline, slant, size, margin, pressure, spacing, zones and, loop of ‘e’ and the distance between title(dot) and stem of ‘i’ are considered for developing the handwriting analysis system.[13] This is what makes each handwriting unique.

II. PROBLEM STATEMENT

Nearly 150 million Indians are in need of active interventions to help with their mental health problems out of which only 41% Indian youth are in favour of seeking support.

Due to the lack of user-friendly methods for analysing an individual’s handwriting for psychological and psychiatric evaluation, the proposed system can be used to extract features from handwriting and use it to draw conclusions about a person’s mental health status.

III. LITERATURE SURVEY

The summary of some of the papers we referred to are shown below in tabular format.

In paper [10], the study is done to see whether graphological analysis is a reliable factor for identifying the symptoms of major depressive diseases. In graphology the handwriting of a person is studied to get some insights from it. According to the methodology used in this paper, 13 handwritten samples of people who are affected by major mental disorders are collected. Few participants are selected and divided into two groups such as: Clinical Sample and Control Group, and evaluation of agreement among graphologists and between the psychological assessment and the judges’ ratings is done. Satisfactory results are found.

In this paper [11], the data collection of DASS is taken from 1771 participants, who are all adults. It involves the process of Statistical analysis to form some competitive models of DASS. These models are later parameterized as model 1a, 2a-2c and 3a-3d. Each of these models measure separate factors like anxiety, stress, etc. The resulting DASS scores are highly influential on Demographic variables. These are used in deriving an optimal DASS model, which is proven to be valid for most of the cases. 
In paper [12], the handwritten text is recognized using the concepts of deep learning. It classifies handwritten text to individual words, to convert the text into digital format. It uses two main methods to complete the task. In the first method, CNN training model is used to classify the words which are derived from handwritten text. In the second method Long-Short term memory networks are used with convolution to build boxes for each character present the word. These characters are then passed to the CNN classifier for classification. At last reconstruction and segmentation takes place.  

The paper [13], gives insights about the outstanding performance of Convolutional Neural Network (CNN) in classification and also explains how they can be improved. Visualization technique is used to analyse the functions of different layers of the CNN classifier.

The paper [14], presents some of the best practices that document analysts can use to do research and get better output from neural networks. The approach followed in this paper is - first, the dataset is created by adding some distorted data to increase the size of the dataset. Second, CNN model is used.

In this paper [15], it consists of the researches on 4 fields like software engineering, web customization, gaming and education. The main idea behind this is to collect all the possibilities of getting emotion recognition methodologies in all these fields. It has analysed every possible scenarios in each of these fields and as a result, gave an idea of the various methods of recognising emotions.

IV. METHODOLOGY

In this system, the handwritten text sample and psychological test results will be used to detect the emotional status of a person. There will be the following sections in the system. They are:

1. Emotion status detection using handwritten text input.
2. Emotion detection is done using psychological tests that provide standards of anxiety, stress and depression.

A. System Description
The proposed system’s methodology or system flow diagram is shown in figure 1. In the first step, a handwritten sample and DASS scale result of the writer will be provided as input to the system. We are taking DASS, because we are mainly focusing on the negative traits. Emotion database will be formed using the DASS result of the writer. Then the input image will undergo pre-processing which involves resizing of image and normalisation. In the next step, the convolutional neural network will be used to train the dataset and then testing takes place. Lastly, a new input sample will be provided to predict the emotion based on the trained model, which will test this input sample and give the result as depression, anxiety, stress or normal.

V. FUTURE SCOPE

The proposed system:

1. Can be extended to work as a language independent assistance.
2. Can be improved by increasing the accuracy of the system by reducing the possible errors.
3. Can be expanded to work on both online and offline samples.

Conclusion

In this paper, the main focus is the use of CNN algorithm for identifying negative emotional traits of a person using his/her handwritten text sample and psychological test results. Since, this paper uses the CNN model for feature extraction and classification of samples to different classes of emotion, it is expected to reduce the training time of the system compared to the previous methods used.

References

[1]Gunes, Hatice, and Massimo Piccardi. \"Bi-modal emotion recognition from expressive face and body gestures.\" Journal of Network and Computer Applications 30, no. 4 (2007). [2]Caridakis, George, Ginevra Castellano, Loic Kessous, Amaryllis Raouzaiou, Lori Malatesta, Stelios Asteriadis, and Kostas Karpouzis. \"Multimodal emotion recognition from expressive faces, body gestures and speech.\" In IFIP International Conference on Artificial Intelligence Applications and Innovations, pp. 375-388. Springer, Boston, MA, 2007. [3]Champa, H. N., and K. R. AnandaKumar. \"Artificial neural network for human behavior prediction through handwriting analysis.\" International Journal of Computer Applications (0975–8887) Volume (2010). [4]S.V. Kedar, D. S. Bormane, Aaditi Dhadwal, Shiwali Alone and Rashi Agarwal. “Automatic Emotion Recognition through Handwriting Analysis: A Review.” International Conference on Computing Communication Control and Automation, 2015. [5]Likforman-Sulem, Laurence, Anna Esposito, Marcos Faundez-Zanuy, St´ephan Cl´emenc¸on and Gennaro Cordasco. . ”EMOTHAW: A novel database for emotional state recognition from handwriting and drawing.” IEEE Transactions on Human-Machine Systems 47, no. 2 (2017). [6]Zhao, Huijuan, Ning Ye, and Ruchuan Wang. \"A Survey on Automatic Emotion Recognition Using Audio Big Data and Deep Learning Architectures.\" In 2018 IEEE 4th International Conference on Big Data Security on Cloud (BigDataSecurity), IEEE International Conference on High Performance and Smart Computing,(HPSC) and IEEE International Conference on Intelligent Data and Security (IDS), pp. 139-142. IEEE, 2018. [7]Jiawen Han, George Chernyshov, Dingding Zheng, Peizhong Gao, Takuji Narumi, Katrin Wolf, Kai Kunze. “Sentiment Pen: Recognizing Emotional Context Based on Handwriting Features.” ACM, 2019. [8]Shreeyash Magar, Bhumika Marathe,Sunny Khambayat, Prem Mamadge. “Handwriting Analysis for Mental Disorder Detection.” International Journal of Engineering Research & Technology (IJERT), 2021. [9]Manogna Pallapothu, Pragati Shinde, Vinish Marito. “Mental Health Analysis Using Handwriting by Generating Writing Prompts.” Publication, year: International Journal of creative research thoughts(IJCRT), 2021. [10]Giannini, Marco, Pietro Pellegrini, Alessio Gori, and Yura Loscalzo. \"Is Graphology Useful in Assessing Major Depression?.\" Psychological reports (2018). [11]Crawford, John R., and Julie D. Henry. ”The Depression Anxiety Stress Scales(DASS): Normative data and latent structure in a large non-clinical sample.” British journal of clinical psychology 42, no. 2 (2003). [12]Balci, Batuhan, Dan Saadati, and Dan Shiferaw. ”HandwrittenText Recognition Using Deep Learning.” CS231n: Convolutional Neural Networks for Visual Recognition, Stanford University, Course Project Report, Spring (2017). [13]Zeiler, Matthew D., and Rob Fergus. \"Visualizing and understanding convolutional networks.\" In European conference on computer vision, pp. 818-833. Springer, Cham, 2014. [14]Simard, Patrice Y., David Steinkraus, and John C. Platt. \"Best practices for convolutional neural networks applied to visual document analysis.\" In Icdar, vol. 3, 2003. [15]Ko?akowska, Agata, Agnieszka Landowska, Mariusz Szwoch, Wioleta Szwoch, and Michal R. Wrobel. \"Emotion recognition and its applications.\" In Human-Computer Systems Interaction: Backgrounds and Applications 3, pp. 51-62. Springer, Cham, 2014.

Copyright

Copyright © 2022 Monica S, Khushi Kumar, Anaga A, Anagha S. This is an open access article distributed under the Creative Commons Attribution License, which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited.