arrow_back Taming Convolution Neural Networks for Image Recognition
Retail Loss Prevention based on Deep Learning
Submitted by Subramanya Mayya (@smayya) on Sunday, 30 April 2017
Items left on the bottom of the shopping cart during checkout is a major source of revenue loss to the retail industry. The Bottom of Basket (BoB) loss or shrinkage as it is called, runs into billions of dollars. Advanced computer vision technology can play an important role in preventing this loss. The presentation covers the design and implementation details of such a solution based on Deep Neural Networks. The talk also touches upon a powerful iterative method called Active Learning which aims to reduce the size of the labelled dataset required to achieve a given performance, thereby reducing annotation cost.
- Introduction - What is the problem we are trying to solve
- Data - What does the data look like
- Approach - Overall design approach and NN architecture
- Data Augmentation - Artificially increase dataset size - how and why
- Evaluation Metrics - Choice of metrics appropriate for the application
- Results - Results and trade-offs
- Active Learning - How to achieve higher accuracy with less labelled data
At minds.ai, I design and implement efficient input pipelines and Deep Neural Networks in TensorFlow for varied applications.