Especially during Covid and lockdown, keeping fit has become a problem. This is where AIWA comes in, it uses visual processing technologies and A.I to guide you through different exercises in a fun way. AIWA also makes sure you are performing the exercises in an optimal way, preventing injury. It comes with many additional features such as posture detection and a blink counter to reduce eye strain. It also comes with a game with motion tracking to make your workout a fun experience.
Maintaining good physical and mental health during lockdown has been a very challenging thing for many people as they wont be able to visit the gym or yoga classes so often and also their social interaction with other people has decreased.
Especially during Covid and lockdown, keeping fit has become a problem. This is where AIWA comes in, it uses visual processing technologies and A.I to guide you through different exercises in a fun way lack of accessibility: a large portion of the common populace does not have access to gyms or gym trainers and therefore prefer to workout and get some exercise on their own. This means they are more prone to injuries, may lose motivation and not reach their end goal. During these trying times of the covid 19 pandemic, this problem is further highlighted. day-to-day applicability: Things like bad posture and eye strain, although seemingly trivial, may have complicated long-term effects. People knowingly or unknowingly overlook these things. This is where AIWA comes in.
We all can agree due to lockdown our physical state has degraded due to lack of exercise and metal health has become worse . This is were AIWA comes in it will help you in keeping fit and tracks your emotional state throughout the day so that you can improve your physical and mental health.
We will be using image processing paired with a trained ML/DL module to count the number of repetitions of an exercise you are doing and tells if you are doing something wrong. For Posture detection we will be calibration the model every time as the user will be sitting in different place so that it is more versatile.
We built this project for NMIT Hackthon 22
Download the zip file or clone the repository
git clone https://github.com/adithya-s-k/AIWA-Web.gitGo to the repository where you have cloned and run app.py
python app.pyThe project is built using Python and Flask.
We are using a mediapipe and we are using the pose estimation model.
By using the pose estimation model ,we will be able to get the coordinates of 32 points on our body as shown below.
Using these coordinates we can calculate the angle, distance between the points.
Thus we can set parameters for different exercises depending on the angle between the joints and calculate the number of reps a person is doing in each exercise as show below
We have also used a pre trained model from Keras.Using this we are able to detect different yoga poses a person is doing as shown below.
Preview of the website
Here is the Project Video This video is of the offline version of this app
- You can workout safely at the comfort of your house
- Current workout apps can show only show images and videos of exercise but AIWA will show you images and videos of exercises and also monitor them using your camera
- It is easier than going to the gym.
- you wont be able to get personal attention while attending yoga classes online.
- Adding a voice assitant to guide you during your workouts and also motivate you to do more.
- Training the yoga ML model with professinals
- Increasing the efficiency by trying out different pose estimation models
Client: HTML ,CSS ,Bootstrap
Server: Flask(Python)
Contributions are always welcome!
You can fork the repository and create a pullrequest for contributing.
Please adhere to this project's code of conduct.
React
Typescript
Django
Laravel
Facebook
Microsoft
Google
Alibaba
D3
Tencent