Scribd
Upload
13 views

AI Based Abnormal Human Detection System

Uploaded by

vigneshbhadri4707
Copyright
© © All Rights Reserved
We take content rights seriously. If you suspect this is your content, claim it here.
Available Formats
Download as PDF, TXT or read online on Scribd
13 views

AI Based Abnormal Human Detection System

Uploaded by

vigneshbhadri4707
Copyright
© © All Rights Reserved
We take content rights seriously. If you suspect this is your content, claim it here.
Available Formats
Download as PDF, TXT or read online on Scribd
You are on page 1/ 3

AI-Based Abnormal Human Detection System

This project aims to build an AI-powered system that can monitor people using a camera

and detect abnormal conditions like seizures or body tremors. Once detected, the system

can alert caregivers or medical professionals.

Key Components:

1. Camera Module:

- The camera will continuously monitor the subject, capturing video or still images.

- You can use a webcam or a more specialized camera (e.g., Raspberry Pi camera module).

2. AI/ML Model:

- Use a deep learning model that can analyze the video feed to detect abnormal behaviors.

- Pose Detection: Utilize models like OpenPose or MediaPipe to track human body keypoints

(joints, limbs, etc.).

- Activity Recognition: A Convolutional Neural Network (CNN) or Recurrent Neural Network (RNN)

can be used for detecting abnormal motions like seizures or shaking.

- Pre-trained Models: Start with pre-trained models for pose estimation and train them further with

specific datasets of abnormal behaviors (shivering, seizures, etc.).

3. Data Collection:

- You may need to collect a dataset of videos showing normal and abnormal behavior (such as

people experiencing tremors or seizures).

- Public datasets: You can look for datasets such as the MobiAct Dataset (which contains

abnormal activity data).


4. Processing Unit:

- For the AI processing, you can use a computer with a GPU, or even a Raspberry Pi with edge AI

accelerators like the Google Coral TPU or NVIDIA Jetson Nano.

5. Alert System:

- Once the AI model detects abnormal behavior, the system should send an alert.

- This can be done through:

- SMS/Email Alerts: Use services like Twilio to send messages.

- App Notification: Create a companion mobile app that notifies family members or healthcare

providers.

- Sound Alarm: A buzzer or siren could also be used for instant alerts.

6. Additional Sensors (Optional):

- You could incorporate additional sensors like accelerometers (in wearable devices) to track

abnormal body movements more accurately.

Tech Stack:

- Programming Languages: Python for AI/ML development.

- Machine Learning Libraries: TensorFlow, Keras, OpenCV for image processing, and MediaPipe for

human pose estimation.

- Hardware: Raspberry Pi with camera module (or any USB camera) and accelerometer if needed

for motion detection.

- Alert System: Twilio API for SMS, Firebase for app notifications, or email alerts.

Workflow:
1. Capture Video: The camera constantly monitors the person.

2. Analyze in Real Time: The AI model processes the video feed to detect abnormal body

movements (seizures, tremors, etc.).

3. Decision Making: If abnormal behavior is detected, the system triggers an alert.

4. Send Alerts: Notifications or alarms are sent out to caregivers/medical professionals.

Challenges:

- Data Collection: Finding enough data for training the AI model, especially for rare conditions like

seizures.

- Real-time Processing: Ensuring the system can detect abnormal behaviors in real-time without too

much delay.

- Accuracy: Balancing between false positives and true detection to avoid unnecessary alerts.

You might also like