Assignment42

Using ArcFace network to get the weights and train a face recognition model
There is 75 data ,51 data to train and 23 data to test the model
Model classifies 4 classes

• 100% accuracy on testing data

Assignment43

Multi Layer Perceptron and Deep learning using keras on four datasets

	Accuracy
Dataset	MLP	Deep Learning
mnist	0.96	0.95
fashion_mnist	0.88	0.81
cfar10	0.40	0.69
cfar100	0.15	0.26

Assignment44

Classification model on four datasets

This classification model classifies 4 objects

• Car 🚗
• Dress 👗
• House 🏠
• Pizza 🍕

	Accuracy	Loss
Train	0.79	0.54
Validation	0.72	0.67
Test	0.86	0.34

Detecting normal people and sheykh people

A deep learning model to detect normal and kheykh people,
also you can use this telegram bot @parisabagherzadeh_bot to do the job
Just send the bot a picture of a sheykh or a normal person

The model used here is VGG16

	Accuracy	Loss
Train	0.99	0.01
Validation	0.98	0.13
Test	0.97	0.12

17 Flower classification

A deep learning model using VGG16 convolution neural net is trained to classify flowers

	Accuracy	Loss
Train	0.98	0.06
Validation	0.90	0.48
Test	0.82	1.06

Real time face mask detector

This is a model based on vgg16 which detects face mask in real time
To train the model this dataset https://www.kaggle.com/ashishjangra27/face-mask-12k-images-dataset is used
To use the model :

• first save the model from Face_Mask_Detection.ipynb file
  
• then run inference.py file

	Accuracy	Loss
Train	0.99	0.01
Validation	0.98	0.55
Test	0.98	0.35

Age Estimation

Automatioc human age estimation based on human facial appearance ,
using ResNet50 convolutional neural network
After saving this model from AgePrediction.ipynb file ,
use the command below in AgePrediction.py file for estimating the age :

python3 AgePrediction.py --input input_file_name.jpg

	Loss
Train	2.90
Validation	5.75

House price prediction with images

• Download the dataset
  To download the house prices dataset you can just clone Ahmed and Moustafa’s GitHub repository
  !git clone https://github.com/emanhamed/Houses-dataset
  

• Train the model
  open up a terminal and execute the following command to train Keras CNN for regression prediction
  python cnn_regression.py --dataset ~/Houses-dataset/Houses\ Dataset/

• Use the model for your own images
  It is four images needed to predict the house price
  Open up a terminal and execute the following command:
  python inference.py --input ./images