- Developed a deep learning model to detect armed persons in video feeds from CCTV cams
- Testing various SOTA models like ViT, CLIP, BLIP, OWL-ViT …
- Developed a sub module for predicting weapon type
- Managed a team of 4 junior/middle software developer
- Created dataset of short CCTV video from real incidents
- Explored bunch of scientific articles on CV field
More info on the project page.
Skin photo analyzer for Derma.ru network of clinics.
- Developed a model for distinguishing psoriasis from another skin diseases
- Developer model for predicting PUVA therapy effectiveness on certain patient
- Developed architecture of telemedicine service for dermatologist
- Developed an data marking web-app for medical images
CNN DL Docker grpc ONNX PHP Laravel
"The purpose of the course is to provide graduate students of various faculties of Lomonosov Moscow State University with basic knowledge of programming and mathematics with the opportunity to use artificial neural network methods for big data analysis in their scientific research."
- Developed a course on deep learning and the Pytorch framework
- Created a set of jupyter notebooks with course materials and assignments
- Advising and assisting graduate students with machine learning research
- Lecturing and recording a video on machine learning
ML DL CNN RL Transfer learning Metric Learning Python Pytorch Numpy
Videoanalytics module detecting the state of the door: open or closed. The module had to work on doors of different types, including glass and sliding doors, as well as to react to half-open doors and ignore people and other objects partially overlapping the doorway.
To solve the problem we used a monocular MiDaS depth detector. It was used to analyze the depth changes on the edges of the doorway.
Depth estimation OpenCV Doccker grpc
Boundary recognition of certain types of documents: passports, ID card, ticket ...
It was required to define clear document boundaries, assuming that each boundary can be defined by a straight line. That is the document is described by 4 points instead of 2 as with the normal detection. The problem was solved by segmenting the photo and then post-processing it.
The model was ported from Pytorch to C++ to run in the product environment.
DL UNet Segmentation OpenCV ONNX C++ Docker
Bicycle detection and recognition based on the video from surveillance cameras*
- Built a real-time video processing pipeline that utilized deep learning techniques on GPU servers
- Developed a cascade of deep learning model to detect and describe images of bicycle from various sources
- Trained the object detector(bicycle) for low-res, and night videos
- Implemented and tested object tracking algorithm
- Built a web service to track movement of certain bicycle on town map
- Developed data-markup system and created dataset of bicycle parts (~30K)
- Optimized computer vision models for faster execution with optimization techniques such as distilling, and quantization.
- Created an API server for external bike recognition
Python Pytorch OpenCV CNN DL ResNet* YOLOv3 Faster-RCNN ONNX Docker TensorRT PHP Laravel grpc Prometheus
Two-way data exchange with GIS of housing and communal services in synchronous and asynchronous modes.
Connecting to third-party services, importing readings of metering devices. Integration with FIAS base.
The code is closed. Fragments on request
One of the largest Russian off-road sites. Commercial successful project, more than 1,500 visitors per day (2015).
In addition to the site itself, its own CMS, CRM, stock parser, and report generation mechanism were created. Integration with the API of IP telephony service providers and SMS mailings has been carried out.
Participated in the development of a universal Logo-based project environment for elementary and preschool education. This software has been widely used in Russian and Moscow schools.
- Development of components in Ninja and C++
- Development under MS Windows
- Development of networking protocols