Introduction

Traditional presentation methods often rely on physical devices like clickers, keyboards, or mouse controls, which can disrupt the flow of a presentation and limit interactivity. These methods can be especially cumbersome in virtual or online meetings, where presenters may struggle to manage slides, videos, and other content seamlessly.

The introduction of hand gesture control in this presentation system addresses these challenges by providing an intuitive, touchless way to control and navigate through presentations. Hand gestures enhance the natural flow of the presentation, allowing presenters to focus more on their content and audience rather than managing devices.

This innovation is particularly significant in the context of online meetings where touchless, gesture-based interaction creates a more engaging and dynamic experience for both presenters and viewers. It allows users to present with ease, whether in corporate environments, classroom settings, or any virtual setting, making it easier to deliver smooth, professional presentations.

Ai Part Demo

Project Description

The AI-Powered Presentation System is an advanced, fully web-based application designed to redefine how presentations are created, controlled, and managed. Built on the versatile Flask framework, the system integrates artificial intelligence with the powerful CVZone library to enable seamless interaction through hand gestures.

This project introduces a natural, intuitive way to control presentations without traditional devices such as clickers or keyboards. By recognizing specific hand gestures in real-time, users can perform actions like:

• Navigating slides forward or backward.
• Starting or ending a presentation effortlessly.
• Controlling videos embedded in slides.
  In addition to gesture control, the system offers comprehensive functionality for managing presentations and user interactions. Key features include:
• User Management: Supports user login, registration, and profile management.
• Content Management: Allows users to upload, manage, and organize presentations and videos.
• SQL Database Integration: Ensures secure and efficient storage of user data, presentation files, and activity logs.
• Web-Based Accessibility: Operates fully through a browser, making it easy to access on any device without the need for additional software or specialized hardware.
  This system is designed to cater to a wide range of users, from educators and corporate professionals to content creators, providing a streamlined and interactive way to deliver engaging presentations.

Gesture Recognition and Control with AI

MediaPipe, developed by Google, is used primarily for real-time processing of multimedia data, including video, audio, and sensor data. Some key features of MediaPipe include:

• Hand Tracking: MediaPipe’s hand tracking model can detect up to 21 key points on a hand, enabling it to recognize detailed gestures like pinch, swipe, or fist. This makes it ideal for applications like gesture-based control in presentations.
• Pose Estimation: MediaPipe provides accurate tracking of the human body in space, which could also be used in other types of gesture recognition systems.
• Face Detection & Landmarks: Detects facial features for use in applications like emotion recognition or virtual avatars.
• Real-time Performance: Optimized to run in real-time across various devices and platforms, including smartphones and web applications.

Database integration

The AI-powered presentation system utilizes a MySQL database to efficiently manage user data, presentation content, and actions. This ensures that the system remains highly functional and scalable while securely storing and retrieving data. The database handles all user-related and presentation-related actions, from logging in and managing profiles to uploading content.

user = "root"
pin = ""
host = "localhost"
db_name = "Presnterx"

# Configuring database URI
app.config['SQLALCHEMY_DATABASE_URI'] = f"mysql+pymysql://{user}:{pin}@{host}/{db_name}"
 
# Disable modification tracking
app.config['SQLALCHEMY_TRACK_MODIFICATIONS'] = False
db = SQLAlchemy(app)
Migrate(app,db)

Project Features

The AI-powered presentation system offers a robust, interactive, and efficient solution for managing and delivering presentations. It leverages advanced technologies like AI and hand gesture recognition,
providing users with an innovative way to create, customize, and control presentations.
1. User Registration and Login:

• Users can securely register and log into the system.

• Once registered, they can access all features, ensuring a personalized and secure environment for managing their presentations.
  2. Creating and Uploading Presentations:

• Users can create new presentations by adding presentation details and uploading their content (slides, videos, or images).

• Uploaded content is securely stored in the MySQL database, ensuring accessibility and organization.
  3. Customizable Presentation Settings:
  Each presentation includes a set of customizable settings that are stored in the database and applied during the session to enhance interactivity and user experience. These settings include:

• Hand Selection:
  The user can define which hand (left or right) will be used for gesture control during the presentation. This flexibility ensures comfort and adaptability for all presenters.

• Presenter Visibility:
  The presenter’s live feed is displayed in a small, interactive frame overlayed on the main presentation slide. This feature fosters better interaction and engagement with the audience by maintaining the presenter’s presence throughout the session.

• Color Preferences:
  Users can customize the color themes used during the presentation, ensuring alignment with their branding or style preferences.

• Presentation Mode:
  Users can choose how to navigate slides based on their preferred interaction style:

• Hand Movement: Navigate slides using full hand gestures.

• Finger Gestures: Use precise finger movements for slide control.
  4. Gesture-Based Control:
  The system incorporates the CVZone library for real-time hand gesture recognition, enabling users to control presentations effortlessly. Key gesture controls include:

• Slide navigation (next/previous).

• Starting and ending the presentation session.

• Managing embedded video playback.

Solution

The AI-powered presentation system showcases the magic of artificial intelligence and its role in driving digital transformation. By integrating advanced gesture recognition, this system revolutionizes traditional presentation methods, enabling seamless and intuitive control without relying on a mouse or keyboard. After setting up default preferences, users can harness the full potential of AI to interact with their presentations effortlessly.
Key AI-Driven Features

1. Slide Navigation with Finger or Hand Movement:

• Users can navigate between slides simply by swiping their hand or moving a finger in the air.
• This hands-free control adds convenience and ensures a smooth presentation flow.

2. Spotlight on Specific Areas Using Two Fingers:

• Highlight specific areas on a slide by bringing two fingers together and directing them toward the desired spot.
• This feature mimics the functionality of a laser pointer, improving audience focus on critical content.
  3.On-Screen Writing with One Finger
• Presenters can write or draw on the presentation screen using a single finger.
• This feature is ideal for real-time annotations or emphasizing key points.
  4.Show and Hide Presenter Frame Using Five Fingers:
• Display or hide the presenter's live video feed by opening or closing five fingers.
• This quick and interactive feature adds versatility and personalization to the presentation.
  5.Close the Presentation with One Hand
• A specific gesture using one hand allows users to end the presentation without the need for external devices.
• This feature simplifies the session wrap-up process and enhances user control.
  6.Frame Mode Flexibility Based on Hand Type
• The system dynamically adjusts the frame mode based on the hand type (left or right), providing flexibility for users.
• This adaptability ensures comfort and ease of use for both left-handed and right-handed presenters.
  7.Erase On-Screen Annotations Using Three Fingers
• Users can clear their on-screen writings or drawings by performing a three-finger swipe.
• This feature ensures a clean workspace for continuing the presentation.

code snippet

To ensure seamless operation and customization, the system begins by defining a set of default variables. These variables serve as the foundation for configuring the presentation environment and allow users to personalize their experience further. The default settings are stored in the MySQL database, enabling persistence and adaptability for all sessions.

#######################################detectors#############################################################################
detector = HandDetector(staticMode=False,maxHands=1,modelComplexity=1,detectionCon=0.5, minTrackCon=0.5)   
face_detector = FaceDetector(minDetectionCon=0.5, modelSelection=0)
############################################################################################################################
# Initialize hand, face, and other detectors
detector = HandDetector(staticMode=False, maxHands=1, modelComplexity=1, detectionCon=0.5, minTrackCon=0.5)
face_detector = FaceDetector(minDetectionCon=0.5, modelSelection=0)
from myproject import app

# Define a function to generate frames for streaming
def generate_frames(id):

    global imageNumber, faceWhiteImageMode, handWiteImageMode, handType, y_axix, buttonPressed, buttonDelay, buttonCounter, presnterImageMode, presnterDelay, PresneterCounter, pointerColor, annotationsStart, annotationNumber, previous_x, presntionSlidingMode, gray_color, closeCamera, closeCameraCounter, closeCameraDelay, annotations, fingers
    with app.app_context():
        #intiate db for context
        db   
        presntaion = Presentaions.query.filter_by(id=id, deleted=1).first()
        if presntaion:
            presntaion_id = presntaion.id
            owner = presntaion.user_id
            images = PresntaionImages.query.filter_by(pres_author=presntaion, deleted=1).all()
            image_names = [image.image_name for image in images]
            folderPath = current_app.root_path + f'/static/pres_images/{owner}/{presntaion_id}/'
            presnetaionPath = sorted(os.listdir(folderPath), key=len) 
    while True:
        success, frame = cap.read()
        if not success:
            break
        frame = cv2.flip(frame, 1)
        frame = cv2.resize(frame, (wCam, hCam))
        currentImagePath = os.path.join(folderPath, presnetaionPath[imageNumber])
        currentImage = cv2.imread(currentImagePath)
        cv2.line(frame, (0, y_axix), (wCam, y_axix), (0, 255, 0), 10)
        hands, _ = detector.findHands(frame, draw=True, flipType=True)
        if hands and buttonPressed is False:
            hand = hands[0]
            handList = hand['lmList']
            cx, cy = hand['center']
            indexFinger = hand['lmList'][8][0], hand['lmList'][8][1]
            if hand['type'] == 'Left':
                xVal = int(np.interp(handList[8][0], [wCam // 2, wCam - 200], [0, wCam]))
                yVal = int(np.interp(handList[8][1], [200, hCam - 200], [0, hCam]))
 
            indexFinger = xVal, yVal
            fingers = detector.fingersUp(hand)
            if cy <= y_axix:
                if previous_x is not None:
                    if presntionSlidingMode == 0:
                        if fingers == [0, 1, 1, 1, 1]:
                            if cx - previous_x < -50:
                                annotationsStart = False
                                if imageNumber < len(presnetaionPath) - 1:
                                    buttonPressed = True
                                    annotations = [[]]
                                    annotationNumber = 0
                                    imageNumber += 1
                                    previous_x = None
                      
                    else:
                        if fingers == [1, 0, 0, 0, 0]:
                            annotationsStart = False
                            if imageNumber > 0:
                                buttonPressed = True
                                annotations = [[]]
                                annotationNumber = 0
                                imageNumber -= 1
                        if fingers == [0, 0, 0, 0, 1]:
                            if imageNumber < len(presnetaionPath) - 1:
                                buttonPressed = True
                                annotations = [[]]
                                annotationNumber = 0
                                imageNumber += 1
               
            elif cy >= y_axix:
                if fingers == [0, 0, 0, 0, 0] and PresneterCounter == 0 and presnterImageMode is True:
                    PresneterCounter = 1
                if fingers == [1, 1, 1, 1, 1] and PresneterCounter == 0 and presnterImageMode is False:
                    PresneterCounter = 1
            if fingers == [0, 1, 1, 0, 0]:
                cv2.circle(currentImage, indexFinger, 10, pointerColor, cv2.FILLED)
                cv2.circle(currentImage, indexFinger, 50, gray_color, 30)
        
                
            if fingers == [0, 0, 1, 1, 1]:
                if annotations:
                    if annotationNumber >= 0:
                        annotations.pop(-1)
                        annotationNumber -= 1
                        buttonPressed = True
        else:
            annotationsStart = False
            previous_x = None
        if PresneterCounter > 0 and presnterImageMode is True:
            PresneterCounter += 1
            if fingers != [0, 0, 0, 0, 0]:
                PresneterCounter = 0
            if PresneterCounter > presnterDelay:
                presnterImageMode = False
                PresneterCounter = 0
        
        if closeCameraCounter > 0 and fingers == [0, 0, 0, 0, 0]:
            closeCameraCounter += 1
            if closeCameraCounter > 15:
                cv2.circle(currentImage, (100, 100), 30, pointerColor, cv2.FILLED)
                cv2.putText(currentImage, str(closeCameraCounter - 1), (80, 110), cv2.FONT_HERSHEY_SIMPLEX, 1, gray_color, 2, cv2.FILLED)
            if closeCameraCounter > closeCameraDelay:
                cap.release()
                closeCameraCounter = 0
        else:
            closeCamera = False
            closeCameraCounter = 0
        if buttonPressed:
            buttonCounter += 1
            if buttonCounter > buttonDelay:
                buttonCounter = 0
                buttonPressed = False
        for i in range(len(annotations)):
            for j in range(len(annotations[i])):
                if j != 0:
                    cv2.line(currentImage, annotations[i][j - 1], annotations[i][j], pointerColor, 10)
        
        if PresneterCounter > 0 and presnterImageMode is True:
            PresneterCounter += 1
            if fingers != [0, 0, 0, 0, 0]:
                PresneterCounter = 0
            if PresneterCounter > presnterDelay:
                presnterImageMode = False
                PresneterCounter = 0
     
        ret, buffer = cv2.imencode('.jpg', currentImage)
        frame = buffer.tobytes()
        yield (b'--frame\r\n' b'Content-Type: image/jpeg\r\n\r\n' + frame + b'\r\n')

Full Project Demo

Conclusion

The AI-powered presentation system represents a significant leap forward in how presentations are created, managed, and delivered. By combining the power of artificial intelligence, intuitive gesture control, and robust web-based technologies, the system eliminates the reliance on traditional tools like mice and keyboards, offering an innovative, hands-free experience.

Key highlights of the project include:

1-Seamless Interaction: Users can control their presentations with simple hand and finger gestures, making the process natural and engaging.
2-Personalization: Customizable settings allow presenters to tailor the experience to their specific needs, from hand preferences to color themes.
3-Enhanced Engagement: Features like presenter visibility and on-screen annotations foster better interaction with the audience.
4-Technological Integration: The use of the CVZone library for gesture recognition, combined with a MySQL database for data management, ensures reliability, scalability, and efficiency.
This system is designed to cater to a diverse range of users, including educators, business professionals, and general audiences, offering flexibility and adaptability for any online meeting or presentation setting. It embodies the transformative potential of digital innovation, simplifying presentation workflows while enhancing user experience.

By bridging the gap between advanced AI capabilities and user-friendly interfaces, this project demonstrates how technology can empower individuals to communicate more effectively and efficiently. It sets a new standard for interactive presentations, making them smarter, more accessible, and future-ready.