There is a ai innovation competition in our university on physiotherapy day I have selected topic cardiopulmonary Suggest me ideas it can be related to breathing exercises and ai innovations already done and which can be done on this
Reading File
Finding Sources
Searching PubMed
artificial intelligence AND pulmonary rehabilitation
Searching the Web
AI breathing exercise physiotherapy pulmonary rehabilitation real time feedback app research 2025
For a university AI innovation competition, choose a project that solves a physiotherapy problem visibly: incorrect breathing pattern, poor home-exercise adherence, and lack of real-time feedback.
Best project idea: BreatheWise AI
An AI-powered breathing-exercise coach for cardiopulmonary physiotherapy
What it does
Using a phone camera, microphone, or low-cost wearable sensor, the system:
- Detects chest and abdominal movement during breathing.
- Measures breathing rate and inhale:exhale ratio.
- Identifies common errors, such as:
- Shoulder elevation and accessory-muscle overuse
- Fast/shallow breathing
- Breath-holding
- Poor diaphragmatic movement
- Incorrect pursed-lip breathing timing
- Gives live feedback:
“Breathe out more slowly through pursed lips” or “Try to expand your abdomen, not your shoulders.” - Tracks exercise completion, symptoms, perceived exertion, and progress over time.
- Flags unsafe readings or symptoms for physiotherapist review, but does not diagnose disease.
Exercises it can guide
- Diaphragmatic breathing
- Pursed-lip breathing
- Thoracic expansion exercises
- Segmental breathing
- Inspiratory muscle-training sessions
- Paced breathing for anxiety-related dyspnea or cardiac rehabilitation
- Post-operative deep-breathing practice and incentive-spirometry adherence
Target users
- COPD and asthma rehabilitation patients
- Post-COVID respiratory rehabilitation
- Post-operative thoracic/abdominal surgery patients
- Cardiac rehabilitation patients
- Older adults with reduced exercise tolerance
- People doing home-based pulmonary rehabilitation
Breathing exercises and pulmonary rehabilitation are established components of recovery in respiratory and surgical settings, including aerobic exercise, deep breathing, and inspiratory-muscle training. Miller’s Anesthesia, 10e, pp. 832-846.
Why it is a strong competition project
Most students will make a chatbot. Your innovation should show AI sensing plus real-time correction, which is more physiotherapy-specific.
Your key message:
“Existing apps tell patients what exercise to do. BreatheWise AI checks whether they are doing it correctly.”
Existing AI innovations: what has already been done
| Existing area | Current examples | Gap you can address |
|---|---|---|
| Wearable respiratory monitoring | Sensors and AI can recognize breathing patterns from body movement signals. A recent system combined an IMU and flex sensor for real-time AI breathing-pattern recognition. | Often expensive, research-only, and not designed around physiotherapy exercises. |
| Device-guided pulmonary rehabilitation | Respiratory-training devices can display breathing curves and give real-time feedback. A 2026 trial reported improved short-term outcomes in hospitalized AECOPD patients using a multiparameter training device. | Limited accessibility, device dependence, and less focus on home practice. |
| Tele-rehabilitation | Video sessions, wearables, and apps monitor adherence and exercise intensity remotely. | A therapist cannot continuously watch every home session. |
| AI in respiratory medicine | AI is widely applied to imaging, pulmonary hypertension, lung-function interpretation, and cough/audio analysis. | Less focus on correcting the quality of breathing exercise performed by the patient. |
The opportunity is to make the system low-cost, home-based, physiotherapist-supervised, and exercise-specific. A recent device-guided rehabilitation study supports the value of real-time respiratory feedback, but your concept can make feedback more accessible through a smartphone.
Three versions, depending on your resources
1. Easy prototype: AI Breathing Coach with Phone Camera
Best if you only have a few days or limited hardware.
- Use a smartphone camera positioned in front of the chest/abdomen.
- Apply pose-estimation AI to track shoulder, chest, and abdominal landmarks.
- Display a breathing waveform.
- Classify the session as:
- Correct diaphragmatic breathing
- Chest-dominant breathing
- Shoulder-dominant breathing
- Irregular or rapid breathing
- Provide voice feedback and a score out of 100.
Demo: Ask one participant to breathe normally with shoulder movement, then demonstrate correct diaphragmatic breathing. Your dashboard detects the difference.
Tools: MediaPipe Pose, Python, Streamlit, OpenCV, TensorFlow Lite.
2. Better prototype: Smart Breathing Belt
Best balance of innovation and feasibility.
Components:
- Stretch sensor or force-sensitive resistor around lower ribs/abdomen
- ESP32 or Arduino
- Optional pulse-oximeter sensor
- Bluetooth connection to a mobile/web app
- AI model to classify respiratory pattern
Innovation: Compare upper-chest expansion with abdominal expansion. The app encourages diaphragmatic breathing and detects paradoxical or shallow patterns.
Output screen:
- Respiratory rate
- Inhalation and exhalation duration
- Abdominal-to-chest movement ratio
- Breathing regularity
- Exercise adherence
- “Technique quality” score
3. Advanced, high-impact idea: AI Physiotherapy Copilot for Pulmonary Rehabilitation
This is a full platform rather than just an exercise counter.
It combines:
- Phone-camera movement analysis
- Wearable breathing sensor data
- SpO₂ and heart-rate input, if available
- Patient-rated dyspnea, such as Borg scale
- AI-based personalized progression
For example:
“The patient completed three sessions correctly, respiratory rate recovered faster than last week, and dyspnea rating improved. Progress from 5 to 7 minutes of paced pursed-lip breathing, subject to physiotherapist approval.”
Safety rules: If low oxygen saturation, chest pain, severe breathlessness, dizziness, or unusual heart-rate data are entered, it stops exercise advice and tells the user to contact a clinician/emergency service as appropriate.
Do not claim that it can diagnose COPD, heart failure, pneumonia, or pulmonary embolism. Frame it as a clinical decision-support and exercise-feedback tool.
Other original project ideas
A. AI Incentive Spirometry Adherence Monitor
A common problem is that patients are prescribed deep breathing or incentive spirometry but perform it inconsistently or incorrectly.
Your system could:
- Use a camera or low-cost pressure sensor to detect usage.
- Count correct repetitions.
- Track inspiratory hold time.
- Give gamified targets.
- Create a physiotherapist dashboard for adherence.
Good target group: post-operative patients at risk of atelectasis.
B. AI Dyspnea-Aware Exercise Pacing System
Patients with COPD or cardiac conditions may stop too early or push too hard.
The app combines:
- Respiratory rate
- Heart rate
- Pulse oximetry, if available
- Borg dyspnea score
- Speech analysis, such as whether the person can speak a full sentence
It then recommends:
- Continue
- Slow down
- Rest with pursed-lip breathing
- Contact clinician according to safety protocol
C. Gamified Breathing Rehabilitation for Children
Make breathing exercises a game:
- Blow virtual bubbles through prolonged exhalation
- Move a bird or balloon by breathing slowly
- Earn points for correct inhale-exhale timing
- Parent and physiotherapist progress dashboard
This is visually attractive for competition judges and can be useful in pediatric respiratory physiotherapy, but you must emphasize that it is an adjunct to therapist-guided care.
D. AI Cough and Breathing Diary
The user records a short daily audio clip and symptom log.
AI can:
- Count cough events
- Track breathing rate from audio
- Detect change from the individual’s baseline
- Generate a simple trend graph for therapist review
Keep its purpose as monitoring and triage support, not diagnosis. Audio-based respiratory AI is an active area of research, as summarized in a recent respiratory AI review.
E. Posture-Breathing AI Analyzer
Poor forward-head and slumped posture can reduce chest expansion and alter breathing mechanics.
Using a camera, it:
- Detects posture during breathing exercises
- Measures shoulder elevation and thoracic movement
- Gives correction: “Sit upright,” “Relax shoulders,” “Expand lower ribs”
- Combines postural correction with diaphragmatic breathing training
This is easy to demonstrate live.
Recommended final title and pitch
Title
BreatheWise AI: Real-Time Intelligent Feedback for Safe Home Cardiopulmonary Physiotherapy
30-second pitch
“BreatheWise AI is a low-cost smartphone and wearable-supported system that helps patients perform breathing exercises correctly at home. It uses AI to analyze chest, shoulder, and abdominal movement, detects poor diaphragmatic breathing and incorrect pursed-lip breathing, and provides real-time corrective feedback. It allows physiotherapists to monitor adherence and progress remotely, making cardiopulmonary rehabilitation more accessible, personalized, and safer.”
What to show judges
- Problem: Patients often do breathing exercises incorrectly and physiotherapists cannot supervise every home session.
- Solution: AI detects breathing pattern and gives immediate correction.
- Live demo: Correct versus incorrect diaphragmatic breathing.
- Dashboard: respiratory rate, exercise quality, adherence, weekly progress.
- Clinical safety: not for diagnosis, uses therapist-set limits and symptom red flags.
- Impact: affordable remote rehabilitation for COPD, cardiac rehabilitation, post-operative and rural patients.
My recommendation
Build BreatheWise AI using smartphone camera posture analysis plus a simple smart breathing belt. It is novel enough for an innovation competition, visibly demonstrates AI, relates directly to cardiopulmonary physiotherapy, and can be developed as a prototype without needing a hospital-grade device.
This is a shared conversation. Sign in to Orris to start your own chat.