Cornell Virtual Workshop > Building Scalable CNN Models > Image Classification and CNN

Background

Traditional Methods

Traditional image classification methods consist of two steps:

Feature extraction (encoding)
Establishing a connection between features and image labels (decoding)

traditional_method — Traditional method
Image source: [1]

What is Deep Learning?

As we mentioned, traditional approaches first identify the features to solve a problem, then develop methods to extract these features. In contrast, deep learning methods identify the features needed to solve a problem using the presented data.

References:

[1] Centiner. URL: https://github.com/NHERI-SimCenter/SimCenter_DesignSafe_ML_2022/blob/main/presentations/day3/MLTraining2022_ImageClassification.pdf

[2] Mahapatra. URL: https://towardsdatascience.com/why-deep-learning-is-needed-over-traditional-machine-learning-1b6a99177063

Back

© Chishiki-AI | Cornell University | Center for Advanced Computing | Copyright Statement | Access Statement
CVW material development is supported by NSF OAC awards 1854828, 2321040, 2323116 (UT Austin) and 2005506 (Indiana University)