Autism spectrum disorders (ASD) comprise a range of neurodevelopmental disorders characterized by limitations in social interaction, communication, and repetitive behaviors. Repetitive hand motions are a common stimulus behavior for individuals with autism, serving as a nonverbal communication cue and self-regulatory mechanism. The project examined the signature hand movements of individuals with autism and used machine learning algorithms focused on identifying three specific repetitive behaviors in children with autism: arm flapping, hand clapping, and finger rubbing. The aim is to support early and accurate ASD diagnosis. Through advanced video analytics and pattern recognition techniques, we strive to contribute to a deeper understanding of ASD behavioral markers.
- matplotlib==3.8.2
- numpy==1.24.1
- opencv-python==4.8.1.78
- Pillow==9.3.0
- six==1.16.0
- tensorboard==2.15.1
- tensorboard-data-server==0.7.2
- torch==2.1.2+cu121
- torchvision==0.16.2+cu121
C:.
│ clip_cap.py
│ main_for_ori_model.py
│ main_for_potatonet.py
│ ori_model.py
│ potato_model.py
│ README.md
│ vid2fram.py
├─data
│ ├─label
│ ├─out
│ └─vid
├─model
│ ├─ori
│ │ PoseModel_whole.pt
│ │
│ └─PotatoNet
│ PoseModel_epoch60.pt
└─runs
├─pose_model_20240112-133345
│ events.out.tfevents.1705037625.albyq.86228.0
│
└─pose_model_20240112-145836
events.out.tfevents.1705042716.albyq.95204.0
