Behavior Analyzer

Behavior Analyzer is a web-based application, which helps the user to observe the behavior of persons as well as texts or tweets. The person's behavior detection part works on Digital Image Processing (DIP), in which the different attributes of a person's face are observed by using facial expression

Project Title

Behavior Analyzer

Project Area of Specialization

Artificial Intelligence

Project Summary

Behavior Analyzer is a web-based application, which helps the user to observe the behavior of persons as well as texts or tweets. The person's behavior detection part works on Digital Image Processing (DIP), in which the different attributes of a person's face are observed by using facial expression measurements and then the person's behavior is categorized into 5 expressions like neutral, happy, sad, angry, and surprised. On the basis of facial expression, the profile of that person is also ranked and the expressions are graphed with respect to time. The text behavior detection part works on the basis of sentimental analysis. Which observes different texts, articles, paragraphs, and tweets and predicts the impact of the text either it creates a positive or negative impact on the reader.

Project Objectives

• To input image(s) (5 Max) for expression checking
• To categorize the inputted image as neutral, happy, sad, angry, or surprised.
• To input textual data (in input field or upload a .csv file)
• To perform sentimental analysis on the textual data
• To return the impact of the text (positive or negative)

Project Implementation Method

Training:

            1. Prepare Dataset

            2. Read Image from each file

              2.1. Image read using OpenCV

              2.2. Convert image into Greyscale

              2.3. Detect face using Harrcascade

              2.4. Using D-Lib to extract landmarks of face

              2.5 Save features in .csv file with its expressions

            3. Repeat step 2 until all folders are complete

            4. Read .csv file

            5. Save features in X-variable and expression label in Y-variable

            6. Initialize Regression Model and Logistic Regression

            7. Fit above model with all data

 8. Save model in dump file

Benefits of the Project

On the basis of facial expression, the profile of that person is also ranked and the expressions are graphed with respect to time. The text behavior detection part works on the basis of Sentimental Analysis, which observes different texts, articles, paragraphs, and tweets and predicts the impact of the text either it creates a positive or negative impact on the reader.

As you know that the human interaction is important in everyday life which cannot be ignored. Being in this highly interactive world, being known of each other’s expressions and mood is very much important. The behavior analyzer proved to be helping in the said task. It is simple enough, easy to use, and as platform independent as possible for the users to feel convenient while using it. Moreover, any sort of writing has a certain impact on its reader, that can either be positive or negative. The application provides the facility to rank the impact of a specific text on its reader by performing a sentimental analysis on the text, which will yield the impact of the said text it has on the reader.

Technical Details of Final Deliverable

Accuracy for Facial Expressions Recognizer

• More than 70% accuracy over 30% Data

Accuracy for Sentimental Analysis

• More than 75% accuracy over 30% Data

Final Deliverable of the Project

HW/SW integrated system

Core Industry

IT

Other Industries

Education , Medical , Security

Core Technology

Artificial Intelligence(AI)

Other Technologies

Artificial Intelligence(AI)

Sustainable Development Goals

Good Health and Well-Being for People, Quality Education

Required Resources

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