We are a start-up in India that have experience in machine learning, deep learning, chat bot development, website and app development. We are well equipped to solve your problem. We can train a deep learning model and provide the python code for inference. We have experience in face detection, face recognition, image tagging, image classification etc.
Problem Statement: Given two images of parent and child, find the similarity between parent and child.
Our approach:
1) Detect and get face of a person from an image using face recognition library (deep learning based library). Input image should have only one face of a person.
2) Get feature vector representing the face using face recognition library (deep learning based)
3) Using these feature vectors of parent and child, build a classifier. The classification score gives whether the parent and child are related. The classifier will be built and validated using data downloaded from internet that has father-son, father-daughter, mother-son, mother-daughter image pairs. Since, the amount of available data is only a few hundred images, deep learning based classifier is not recommended, we will develop the classifier using lightgbm or xgboost package. The classifier score will be considered as percentage matching between parent and child.
4) Develop a Python code for inference. Given two images, python code will output the classifier score.
Please contact us, if you have any queries.