Object Detection is Deep learning problem. It has three sections (Object Recognition, Object localization by bound box, object identification). Initially, we will detect an localize the object.
- Object Recognition
- Object Localization
- Object Tracking(TODO)
First, we will use an object recognition algorithm. There are a couple of algorithms and pre-train model. Which can help be used for object recognition?
- HAAR
- YOLO (You Only Look Once)
- SSD (Single-Shot Detection)
In this, I have used SSD pre-train Caffe model for object recognition.
- Tracking is the process of tracking the co-ordinates of moving object.
Pre-requisite.
- Python 3.5
- OpenCV
- Pre-train model like Caffe or TensorFlow