Data-Efficient Deep Learning Method for Image Classification Using Data Augmentation, Focal Cosine Loss, and Ensemble

07/15/2020

∙

In general, sufficient data is essential for the better performance and generalization of deep-learning models. However, lots of limitations(cost, resources, etc.) of data collection leads to lack of enough data in most of the areas. In addition, various domains of each data sources and licenses also lead to difficulties in collection of sufficient data. This situation makes us hard to utilize not only the pre-trained model, but also the external knowledge. Therefore, it is important to leverage small dataset effectively for achieving the better performance. We applied some techniques in three aspects: data, loss function, and prediction to enable training from scratch with less data. With these methods, we obtain high accuracy by leveraging ImageNet data which consist of only 50 images per class. Furthermore, our model is ranked 4th in Visual Inductive Printers for Data-Effective Computer Vision Challenge.

READ FULL TEXT

Data-Efficient Deep Learning Method for Image Classification Using Data Augmentation, Focal Cosine Loss, and Ensemble

Deep Learning on Small Datasets without Pre-Training using Cosine Loss

Training Efficient CNNS: Tweaking the Nuts and Bolts of Neural Networks for Lighter, Faster and Robust Models

K-LITE: Learning Transferable Visual Models with External Knowledge

Occlusions for Effective Data Augmentation in Image Classification

No Data Augmentation? Alternative Regularizations for Effective Training on Small Datasets

A Strong Baseline for the VIPriors Data-Efficient Image Classification Challenge

Integrating domain knowledge: using hierarchies to improve deep classifiers

Data-Efficient Deep Learning Method for Image Classification Using Data Augmentation, Focal Cosine Loss, and Ensemble

Related Research

Deep Learning on Small Datasets without Pre-Training using Cosine Loss

Training Efficient CNNS: Tweaking the Nuts and Bolts of Neural Networks for Lighter, Faster and Robust Models

K-LITE: Learning Transferable Visual Models with External Knowledge

Occlusions for Effective Data Augmentation in Image Classification

No Data Augmentation? Alternative Regularizations for Effective Training on Small Datasets

A Strong Baseline for the VIPriors Data-Efficient Image Classification Challenge

Integrating domain knowledge: using hierarchies to improve deep classifiers