Graduation Project for Bachelor Degree
By Kenessary Koishybay, Nauryzbek Razakhbergenov.
Mentor: Anara Sandygulova. Nazarbayev University
Pose estimation algortihm is based on tensorflow implementation of Realtime Multi-Person Pose Estimation
Requirements:
- Python 2.7
- OpenCV3
- sklearn
- scipy
- imutils
- xgboost
To run my code you need to type:
python -B Main.py <input_video> <output_video>
Here, arguments <input_video> and <output_video> are optional,
and default values can be seen in the code.
Pose estimation is the process of locating body key points.
Pose estimation problem is usually solved by training Deep Learning architectures with annotated datasets such as
MPII or COCO
We didn't have computational power to train on these datasets. Thus, we tried pre-trained model mentioned at the beginning.
Architecture:
Note, that we decided to remove code concerning EWMA in the final version.
Single Data Sample Length:
290 = 2*14*10 (x,y coords of 14 body parts in 10 frames) + 10(indexes of each frame)
For every (N = 10)th frame:
- Open pose features calculated for every tracked humans
- This features is then added to the previous features of the tracks
- If the length of resulting feature vectors of specific tracks is large enough, feature vectors will be converted to data samples
- These data samples is then goes as input to the machine learning algorithm (XGBoost)
- XGBoost classifies activity of each data sample as code.
- Code is then decoded into Activity Labels
If you wan't to train our activity recognition algorithm to increase accuracy, first you should extract suitable data from videos dataset.
We extracted data samples from KTH dataset.
Code is in ExtractData folder.