Effects of Data Enrichment with Image Transformations on the Performance of Deep Networks
This work addresses the problem of handling unexpected image variations in deep networks for super-resolution, but it is incremental as it applies standard transformations to an existing dataset.
The study investigated how enriching training data with six image transformations affects deep network performance in super-resolution, finding that 180-degree rotation yielded the best single transformation results, while a mix of all transformations led to the highest scores.
Images cannot always be expected to come in a certain standard format and orientation. Deep networks need to be trained to take into account unexpected variations in orientation or format. For this purpose, training data should be enriched to include different conditions. In this study, the effects of data enrichment on the performance of deep networks in the super resolution problem were investigated experimentally. A total of six basic image transformations were used for the enrichment procedures. In the experiments, two deep network models were trained with variants of the ILSVRC2012 dataset enriched by these six image transformation processes. Considering a single image transformation, it has been observed that the data enriched with 180 degree rotation provides the best results. The most unsuccessful result was obtained when the models were trained on the enriched data generated by the flip upside down process. Models scored highest when trained with a mix of all transformations.