CanvasInsight

A growing number of depression cases are going unnoticed or untreated which is giving rise to suicide rates in the country and around the world. About 13,000 to 15,000 people take their lives in Pakistan alone every year, and 130,000 to 150,000 are at risk of doing the same.  The World

Project Title

CanvasInsight

Project Area of Specialization

Artificial Intelligence

Project Summary

A growing number of depression cases are going unnoticed or untreated which is giving rise to suicide rates in the country and around the world. About 13,000 to 15,000 people take their lives in Pakistan alone every year, and 130,000 to 150,000 are at risk of doing the same.  The World Health Organization (WHO) estimates that, until the year 2021, depression will be the second largest cause of the global disease burden.

The scale of the problem, mental illness and depression is too magnanimous for psychiatrists alone to cater to the needs of the population across Pakistan as in the country as of 2017, there were only around 300 psychiatrists and around 160 under training and of these 160, a good majority may be “brain-drained” to the western countries. There is a need to address depression cases and predict them upon their onset so as to prevent the negative and harmful effects of the disease on the child and their families.

The detection of negative emotions through daily activities such as writing and drawing can prove extremely useful for promoting wellbeing. The widespread of human-machine interfaces such as tablets makes the collection of text and drawing samples easier. In this context, we present a first of its kind artificial agent and drawing tool which relates the emotional state, depression, to text and drawing with the aim to make accurate assessments that allow for early identification of depression among children and adolescents. The foundation of this concept is drawn from the age old method of psychology known as "Draw-a-Story" published in 1988 by Rawley Silver. The detection is based off of a questionnaire followed by the individual's illustration of different scenarios through storyboarding on a simple mobile application. The user input is then analyzed by the machine learning algorithm on the basis of features, colors and other drawing characteristics after which the system outputs conclusive results which are sent over a secure network to collaborating hospitals and/or institutes for further processing or appointment notifications. The target audience is limited to children only but can be expanded to include other age groups in the future.

Project Objectives

With this project our team aims to deliver to hospitals around Pakistan, digitalized and artificially computerized agents designed and trained to be able to identify depression symptoms using imaging, text analysis and story boarding and in turn diagnose depression and its type.

The proposed solution stands out not just because there are no artificial intelligent agents in hospitals here but also because this greatly eases the process for doctors while also speeding up that process in turn efficiently accommodating more patients. Our methodology relates to Draw-a-story which was published in 1988 that showed how psychologists used storyboarding complemented by drawing in identifying state of mind of the patient. We intend to make an AI powered application that will analyze the patients state of mind through conversation and based on that conversation have them draw up a story to be analyzed to conclude the final results and provide a diagnosis. Through this project we aim to help revolutionize how we diagnose, treat, and monitor depression. We believe technological approaches such as our proposed solution have the potential to expand access and augment the work of clinicians and illustrates the promise of data science, and more specifically machine learning, in mental health.

Project Implementation Method

We  begin by using Digital image processing techniques to clean scanned reports and demographics obtained from hospitals and institutes to extract viable information which will form a basis for training and grooming the artificial intelligent agent. The next step then would be to identify and convert the reports from image to text to be able to feed it to the agent under training to extract specific keywords against demographics and diagnosis.

The system will then be trained to understand the demographics and how they are used to pinpoint several abnormalities. It then in turn uses that understanding to enable the agent to do the same task.

This trained agent will allow the user to draw a scenario as an input drawing and identify abnormalities in the drawing Such as:

• broken lines

• drawing at the edge of the given area

• drawing few living beings

• weapons

• sad faces etc.

Based on these results it will diagnose the person as moderate to severe depressed. All of this working will be held on a mobile application which will send the collected information over to a private cloud server on which the artificial agent will be deployed. From there it will make necessary conclusions based entirely upon the story and art created by an individual. 

The project’s development plan and methodology along with deliverables is listed below

Phase

Inception

Feasibility Study

Requirement Gathering

System Modelling

System Design

Implementation

Testing

Deployment

System Acceptance

Benefits of the Project

Through our project we aim to facilitate the psychiatrists and clinicians in diagnosing depression amongst children and adolescents especially in difficult cases where the condition is concealed. It is hoped that our product can be launched into the market by being installed in various psychiatric departments in hospitals and clinics across the country.

Additionally, the early detection of depression through the app will greatly prevent psychological suffering for the patients and their families by receiving timely treatment from their doctors.

In his talk, Professor Murad Moosa Khan, President of the International Association for Suicide Prevention (IASP) pointed out Pakistan’s need for a viable and national, mental health strategy, one that involves different stakeholders including the government, public and mental health professionals and NGOs to address the growing concern.

We are positive that our project will be well received by professionals, the government and the people alike if it is well executed.

With our product, we aspire to facilitate the little over 300 psychiatrists in the country to overcome the challenges in the timely identifying of depression amongst children and adolescents in Pakistan and help save the many children and their families whose quality of lives are heavily burdened with the negative effects of depression.

More over technological advancements are a major part of many of the crucial developments required for a country to move forward and bring about change. If a psychatrists work could be facilitated by an artificial agent, this would allow them to devote more time in conducting research and exploring futher in the domain of psychology to create and facilitate refined versions of a project like this one so that better help could be provided to the people not just from Pakistan but all over the world.

Depression is a state of mind due to which many people suffer and one major reason of this constant prolonged suffering is lack of help, they feel as if they are stuck with this state of mind  going about their lives thinking there's nothing they can do. We want people to know we understand what they're going through, this is not a product to advertise or make money from, this is simply long due help that needs to be put out in the open so that precious lives can be saved.

Technical Details of Final Deliverable

The final deliverable consists of 2 main components. The first one is a simple mobile application which will be developed using react native working on different IDEs particularly Android studio, atom and sublime text available for users all over Pakistan to sign up with. This application will enable an individual to to go through a few phases of evaluation such as questionnaires, Completing half drawn figures, story boarding, drawing, etc. Each of these phases will be backed up with proper thorough research and professional consultancy. Such as the questionnaire will consist of a set of questions approved by authenticated psychatrists/psychologists that can sum up the state of problem and characters and/or items involved. The next phase will be entirely dependent upon the answers to these questions, this will make the experience personalized towards the individual being evaluated. How will this be accomplished? 
That brings us to the next major component which is an artificial intelligent agent built and trained on python working with libraries such as Keeras, ImageAI, OpenCV, etc deployed on cloud services provided by Amazon AWS (The purpose of deploying it on a cloud is to minimize bulky size of the app and to better allow collective training of the agent)  to asses each phase of evaluation to draw up the next phase based on the answers of the previous one. It will be trained to evaluate image features, textual keywords, choice of completion, e.t.c. The cloud will also act as a storage for all records of patients to allow incorporation of larger and broader data set for further training of the agent. 
A combination of these two components will provide a whole system which will allow competent medical authorities to detect and diagnose depression cases quicker and more efficiently. (An idea is to also store the records in a blockchain based cloud to allow secure retrival and storage of patient data)

Final Deliverable of the Project

Software System

Type of Industry

IT , Medical , Health

Technologies

Artificial Intelligence(AI), Blockchain, Cloud Infrastructure, Others

Sustainable Development Goals

Good Health and Well-Being for People, Industry, Innovation and Infrastructure, Partnerships to achieve the Goal

Required Resources

If you need this project, please contact me on contact@adikhanofficial.com