Real-Time Object Recognition, Counting and Verification System Development - Computer Vision

Seeking an experienced Computer Vision/AI specialist to develop and optimize a real-time object recognition system with dual functionality:

Object Counting: Accurately count multiple small objects in real-time video/image streams
Verification Matching: Compare detected objects against a reference database of 4,000+ items to determine matches

Technical Requirements:

Multi-feature classification model incorporating: dimensions, color, shape, and text/imprint recognition
Real-time processing capabilities
High accuracy verification against large reference database
Integration with existing Python/PyTorch pipeline
Google Colab/Drive compatibility for training/deployment

Deliverables:

Optimized neural network model(s)
Training pipeline with data augmentation
Inference code for real-time detection
Performance metrics and validation results
Documentation and integration guide

Required Skills:

Expert-level Computer Vision (OpenCV, PyTorch/TensorFlow)
Object detection (YOLO, R-CNN) and classification
OCR/text recognition implementation
Contrastive learning and embedding models
Real-time video processing optimization
Experience with large-scale image databases

Must provide portfolio examples of similar multi-feature object recognition projects.