Utilizing dataset affinity prediction in object detection to assess training data

arXiv:2311.09768v2 Announce Type: replace
Abstract: Data pooling offers various advantages, such as increasing the sample size, improving generalization, reducing sampling bias, and addressing data sparsity and quality, but it is not straightforward and may even be counterproductive. Assessing the effectiveness of pooling datasets in a principled manner is challenging due to the difficulty in estimating the overall information content of individual datasets. Towards this end, we propose incorporating a data source prediction module into standard object detection pipelines. The module …

abstract advantages arxiv bias cs.cv data dataset datasets detection improving object pooling prediction quality sample sampling sparsity training training data type