Proposal :
Topics selected:
• Shape Detection, Recognition
• OCR
• Counting
• Tracking
1. Shape Recognition & Detection
Project #1:
1. Aim of this project:
• Introduce basic ideas of loading an image operations, load, convert, color spaces, Morphologicals and kernels
• Introduce the idea of object detection with different scenarios:
• Real world image
2. Scenario:
• Generate shapes at random areas of the scene
• Overlay and highlight objects in scene for user feedback
• Sort out the objects which one is closes to the left edge
• Project #2:
1. Aim of this project:
• Apply what was learned in project 1 to detect an object in real world ( vehicle
in our context )
• Introduce the idea of features ( corners, edges )
• Introduce the idea of templates
2. Scenario:
• Detect a car in the scene
• Locate car plate in the scene, highlight it
2. OCR
Project #1:
1. Aim of this project:
• Introduce character recognition problems
• Introduce the concept of machine learning in vision
2. Scenario:
• Using ready package to read a character from ideal image, then real world
• Building a Machine Learning Model read a character from ideal image, then real world
• Develop tactics to read a sentence, plate
3. Counting
Project #1:
1. Aim of this project:
▪ Introduce vehicle counting problem
2. Scenario
•Having videos in hand, count the vehicles passing in a specific direction at specific time
4. Tracking
Project #1 :
1. Aim of this project:
• Introduce deeplearning concepts and Single shot detectors
• Introduce motion detection concepts
• Introduce the need to track objects in the scene
2. Scenario:
• Using a trained model detect vehicles in the scene, every N seconds
• Highlight the object if its the same object detected at t and t-n seconds ( Centroid,Kalman, ... )