How to build your own dataset with Neural Network Console?(Part 3)
Deep learning has become vitally important to our daily life. The way that we unlock our phone and the intelligent LED's among the streets both demonstrate the example of implementing AI in terms of image recognition.
There are many open sources on the internet covering AI, including the handwritten digit recognition, but how could we build our own dataset? In this article, I will share an application with you all, the Neural Network Console.
This article will be divided into 4 parts, the first, preparation of the datasets, the second, build the Network model, the third, create the dataset, and lastly, how to evaluate the model.
Creat Data Set
Source Dir: What we have built in part 1
Output Dir: Here should be an empty file for .CSV
Shaping Mode: We will choose trimming instead of resize
Output Colour should set as 3(RBG)
Ratio(%) should be the portioin of training and testing
Training and Evaluation
Learning Curve, this is the result that was shown after the training action. This is a good fit learning curve. A good fit is identified by a training and validation loss that decreases to a point of stability with a minimal gap between the two final loss values.
Confusion Matrix is a useful tool that shows how your model performs, it uses the training and testing set in the create dataset part. The precision is the accuracy of all your data in the training set. This will help you to decide which label or group should be enhanced and which is kept right.
Could the neural network console recognise a face? Yes! It would be interesting using this app as a starter.
Inputting different data sets and the various results generated.