Video Analytics & Vision
Platforms & Embedded systems are becoming intelligent day by day, so vision algorithms play an important role in enabling intelligence in systems. The age old problem has been in achieving accuracies for independent functioning of the intelligent systems and running vision algorithms in real-time.
In addition to vision algorithms, video analytics has been gaining importance in various applications such as surveillance, augmented reality and video conferencing, to name a few.
PathPartner offers development and optimization services for state-of-the-art vision and video analytics algorithms.
Our typical service delivery model follows the listed process:
- Proof-of-concept - Vision algorithm prototype development through openCV, MatLab, Octave, Eigen Library, Point Cloud Library (PCL)
- Data Set Collection and Annotation of the video data
- Tuning of the deep learning algorithm or CNN
- Optimization of the algorithms on ARM, DSPs, GPUs and Vision Cores
We have expertise in the development of following Vision & Video Analytics algorithms:
- Face detection
- People count
- Gender detection & Age group classification
- Gesture detection
- 3D depth map
- Background subtraction
- Tamper detection
- Motion detection and tracking
The PathPartner Advantage
PathPartner specializes in deep learning or CNN and has developed many vision algorithms for industries such as ADAS, Video Surveillance and Digital Signage.