Computer Vision Engineer Job

We are seeking a 'Computer Vision Engineer' to develop AI-driven video analytics solutions for extracting in-depth cricket data from match footage, live broadcasts, and player tracking systems. This role involves leveraging deep learning, pose estimation, and image processing to analyze player movements, shot selection, bowling mechanics, and match dynamics.

Key Responsibilities:

1. Video-Based Cricket Analytics
- Develop computer vision models for player tracking, ball trajectory prediction, and shot classification.
- Extract real-time insights from video streams, including player positions, footwork analysis, and batting/bowling techniques.
-Implement pose estimation (OpenPose, MediaPipe, DeepLabCut) to analyze player body movements.
-Use action recognition models to classify batting strokes, bowling variations, and fielding plays.

2. Deep Learning & Image Processing
- Apply CNNs, Transformers, YOLO, Mask R-CNN, and Optical Flow for cricket action recognition.
- Develop segmentation models to differentiate players, umpires, and key match elements (ball, stumps, pitch zones).
- Use GANs and super-resolution techniques to enhance low-quality video footage.
- Build OCR (Optical Character Recognition) systems to extract scoreboards and match statistics from live broadcasts.

3. Real-Time Video Processing & Streaming
- Implement real-time video processing pipelines using OpenCV, FFmpeg, and GStreamer.
- Work with multi-angle TV footage and drone feeds to enhance match analysis.
- Develop low-latency algorithms for real-time ball speed, swing, and spin detection.

4. Data Collection, Annotation & Model Training
- Build automated annotation tools for labeling cricket match videos.
- Train supervised and unsupervised learning models using cricket datasets.
- Collaborate with sports analysts to validate and refine model predictions.

5. Model Deployment & Performance Optimization
- Deploy computer vision models on edge devices, cloud platforms, and real-time streaming systems.
- Optimize models for low latency and high accuracy in live cricket broadcasts.
- Implement A/B testing and continuous monitoring to improve model performance.

6. Camera Calibration & Video Trigonometry
- Perform camera calibration for distortion correction and real-world coordinate mapping.
- Use homography transformation and perspective correction for multi-angle alignment.
- Apply video trigonometry to calculate player distances, ball speed, angles, and movement trajectories.
- Work with multi-camera synchronization and stereo vision for depth estimation in cricket analytics.

2