How yang20040522/tkuim_project works
RehabAssist is a Flutter rehabilitation app for patients and therapists. Its principal patient workflow is to view a therapist’s plan, perform daily exercises, use on-device pose recognition to assess movement, and review training records and statistics. Therapist capabilities include patient and plan management plus a 3D custom-exercise editor; chat, calls, profiles, and notifications are supporting branches. The README names these app modules, but their Dart sources are absent from the supplied tree, so those links reflect documented capabilities. The sampled ML code describes a separate research/training workflow, not evidence of app runtime integration.
Main entry points: training.py Body training · training Daily training · plan Rehab plans · features.py Body features
Components 34
Patient rehabilitation
- Rehab plans
lib/features/plan - Daily training
lib/features/training - Rehab exercises
lib/features/rehab - Training history
lib/features/history - Training statistics
lib/features/stats - Patient home
lib/features/home
Therapist tools
- Patient management
lib/features/friends - Custom exercises
lib/features/custom_exercise
Communication and account
- Account and profile
lib/features/account - Chat
lib/features/chat - Video calls
lib/features/call - Notifications
lib/features/notification
On-device pose
- Pose inference
lib/features/analysis - Pose measurement
lib/features/pose_measurement
ML research pipeline
- Body features
ml/body_v3/features.py - Body dataset
ml/body_v3/dataset.py - Body models
ml/body_v3/models.py - Body split
ml/body_v3/split.py - Body training
ml/body_v3/training.py - Body parity
ml/body_v3/parity.py - Synthetic samples
ml/body_v3/synthetic.py - Body train entry
ml/train_body.py - Hand features
ml/hand_features.py - Feature schema
ml/feature_schema.py - Hand model training
ml/train.py - Model artifacts
ml/model_artifacts.py
Desktop application shell
- Windows entry
windows/runner/main.cpp - Flutter window
windows/runner/flutter_window.cpp - Win32 window
windows/runner/win32_window.cpp
Other components
- Patient
- Therapist
- REST backend
- Phone camera
- ONNX Runtime
Connections 42
- Patient → Rehab plans: views
- Rehab plans → Daily training: guides
- Patient → Daily training: performs
- Daily training → Pose inference: analyzes with
- Pose inference → Phone camera: reads frames
- Pose inference → ONNX Runtime: runs models
- Daily training → Pose measurement: judges movement
- Daily training → Training history: records sessions
- Patient → Training history: reviews
- Patient → Training statistics: reviews
- Patient → Patient home: opens
- Therapist → Patient management: manages patients
- Therapist → Rehab plans: arranges plans
- Therapist → Custom exercises: edits exercises
- Custom exercises → Rehab plans: provides exercises
- Patient → Account and profile: manages profile
- Patient → Chat: chats with therapist
- Therapist → Chat: chats with patient
- Patient → Video calls: starts calls
- Therapist → Video calls: starts calls
- Patient → Notifications: receives alerts
- Account and profile → REST backend: exchanges account data
- Patient management → REST backend: exchanges patient data
- Rehab plans → REST backend: exchanges plan data
- Chat → REST backend: exchanges messages
- Body dataset → Body features: extracts features
ml/body_v3/dataset.py - Body parity → Body features: checks parity
ml/body_v3/parity.py - Synthetic samples → Body features: builds samples
ml/body_v3/synthetic.py - Body split → Body features: uses schema
ml/body_v3/split.py - Body training → Body features: uses features
ml/body_v3/training.py - Body training → Body models: uses models
ml/body_v3/training.py - Body training → Body split: splits data
ml/body_v3/training.py - Body training → Body dataset: consumes data
ml/body_v3/training.py - Body train entry → Body training: runs training
ml/train_body.py - Body train entry → Body dataset: loads dataset
ml/train_body.py - Body train entry → Body parity: checks parity
ml/train_body.py - Body train entry → Synthetic samples: generates samples
ml/train_body.py - Feature schema → Hand features: uses features
ml/feature_schema.py - Hand model training → Feature schema: builds feature rows
ml/train.py - Hand model training → Model artifacts: trains artifacts
ml/train.py - Windows entry → Flutter window: creates window
windows/runner/main.cpp - Flutter window → Win32 window: extends window