Case Study: Caterpillar achieves faster, automated classifier development and deployment with MathWorks

Caterpillar Uses Big Data, Data Analytics, and Machine and Deep Learning to Build Ground-Truth for Training, Validation, and Deploying Classifiers

Caterpillar partnered with MathWorks to tackle the challenge of creating reliable ground-truth from large volumes of field data for training, validating, and deploying machine- and deep-learning classifiers. The customer needed a scalable, efficient way to label data, reduce human intervention, and speed up design iterations; they used MathWorks products such as MATLAB and related toolboxes (Computer Vision, Image Processing, Statistics and Machine Learning, plus code-generation and deployment tools).

MathWorks helped implement a big-data infrastructure with a web front end for external labelers, a searchable ground-truth database, and an engineer-facing interface for machine learning, visualization, and code generation. The MathWorks solution automated labor-intensive labeling and classifier comparison, scaled to more users and data, and enabled rapid movement from collected data to an improved classifier running on the machine—resulting in faster design iterations and reduced manual effort.

Caterpillar

Larry Mianzo

Caterpillar