IgNet. A Super-precise Convolutional Neural Network
This work addresses the challenge of handling atypical image data for applications in child development or art analysis, but it appears incremental as it builds on existing CNN frameworks.
The paper tackled the problem of analyzing unique, irregular features in children's drawings using a new CNN called IgNet, achieving 100% accuracy in categorical classification and an error of no more than 0.4% in regression for predicting artists' ages.
Convolutional neural networks (CNN) are known to be an effective means to detect and analyze images. Their power is essentially based on the ability to extract out images common features. There exist, however, images involving unique, irregular features or details. Such is a collection of unusual children drawings reflecting the kids imagination and individuality. These drawings were analyzed by means of a CNN constructed by means of Keras-TensorFlow. The same problem - on a significantly higher level - was solved with newly developed family of networks called IgNet that is described in this paper. It proved able to learn by 100 % all the categorical characteristics of the drawings. In the case of a regression task (learning the young artists ages) IgNet performed with an error of no more than 0.4 %. The principles are discussed of IgNet design that made it possible to reach such substantial results with rather simple network topology.