Current Research

Whenever you are working with EKG, EMG, EEG, ECoG, LFP, Doppler, or accelerometer signals they are different, but from an engineering or abstract mathematical viewpoint they are very similar. They are non-stationary, all have noise, and similar analysis tools could be applied to process them.

The ANRY Lab helps medical professionals analyzing biomedical data and extracting biomarkers and other important information. We develop novel algorithms for automatic signal analysis and classification using time and frequency domain as well as machine learning.

For more information:

Data has no value if not processed and interpreted. Huge amounts of numerical data (text, images, signals, statistics) are recorded and stored every day. Unfortunately, allocating humans to interpret all those data is not feasible, and in many cases, there is a big interobserver variance in their interpretations. If you have data, machine learning (or deep learning) can help to extract features of interest and using those features to interpret it. Machine learning is widely used for prediction, signal or image segmentation or recognition, classification, and other tasks.

The ANRY Lab works on machine learning algorithm applications ranging from driver behavior recognition to activity recognition based on brain signals.

For more information:

Computer vision or image analysis is a broad field of research that allows the computer to segment and interpret images, videos, or biomedical volumes, to match (register) different images, or to reconstruct images (getting better resolution and quality). If you need any of the tasks above, ANRY Lab can help you.

The major goal of ANRY Lab is to develop algorithms for computer vision and biomedical image analysis. Here you will find projects on image segmentation, registration, classification, texture analysis and other various applications of computer vision.

For more information:

Whether you track pedestrians, vehicles, missiles, or bacteria, neurons, protein interactions, nanotubes, or need to predict and estimate a state of a dynamic system, you need visual tracking tools to help you.

We have extensive experience with visual tracking algorithm development for civil, military and biomedical applications, where single or multiple targets are detected or selected in the first frame and followed by the algorithm automatically throughout the video. The projects include tracking of target’s centroid, bounding box, or contour. 

For more information:

More information on completed projects: