.
├── assets
│ ├── banner.png
│ ├── marksman-pipes-vid.mov
│ └── pitch/
├── data
│ ├── example-pose-predictions
│ ├── frames
│ ├── multi-subject
│ ├── pose-estimations
│ └── single-subject
├── inferences
│ ├── quickpose_frame_0001.json
│ └── quickpose_frame_0179.json
├── pages
│ ├── 1_The_Marksman_Training_Problem.py
│ ├── 2_Pose_Analysis_Technology.py
│ ├── 3_Maintainable_and_Cost_Effective.py
│ ├── 4_Superhuman_Feedback.py
│ └── 5_Pose_Analysis_Demo_And_Tech.py
├── project-mgmt
│ ├── 1-todo.md
│ ├── 2-in-progress.md
│ └── 3-done.md
├── LICENSE
├── Marksman_Trainer.py
├── README.md
├── army-marksmanship-unit-rifle.pdf
├── army-rifle-and-carbine.pdf
├── pitch-talk-track.md
├── requirements.txt
├── tech-stack.md
└── .gitignore
See tech-stack.md for details on the technologies used.
-
Create a virtual environment:
python -m venv venv
-
Activate the virtual environment:
source venv/bin/activate -
Install requirements:
pip install -r requirements.txt
-
Run the Streamlit app:
streamlit run Marksman_Trainer.py
- The rifle shooting process is detailed on page 5-2 of tc3-22-9.pdf.
- Minute details of the shooting process are described on page 14 of army-marksmanship-unit-rifle.pdf.
- Improving handgun detection through a combination of visual features and body pose-based data
- Reference Githubs:
- Perplexity: Estimate Yearly Hours of Training in the US Army
Based on the army rifle and carbine documents, the shot process consists of three distinct phases:
Pre-shot phase:
- Position
- Natural Point of Aim
- Sight Alignment/Sight Picture
- Hold
Shot phase:
- Refine Aim
- Breathing Control
- Trigger Control
Post-shot phase:
- Follow-through
- Recoil management
- Call the Shot
- Evaluate
This process forms a complete cycle for each shot taken. The functional elements that support this process include stability, aim, control, and movement – all of which work together to produce accurate and precise shots.
[youtube video links] https://www.youtube.com/watch?v=-e5cjxVynEc