Skew Class-balanced Re-weighting for Unbiased Scene Graph Generation
This work addresses the issue of biased predictions in SGG for computer vision applications, but it is incremental as it builds on prior methods to better balance performance trade-offs.
The paper tackles the problem of unbiased predicate prediction in scene graph generation (SGG) caused by long-tailed distributions, proposing a Skew Class-balanced Re-weighting (SCR) loss function that improves trade-offs between majority and minority predicate performances, with experiments on Visual Genome and Open Image datasets showing its effectiveness.
An unbiased scene graph generation (SGG) algorithm referred to as Skew Class-balanced Re-weighting (SCR) is proposed for considering the unbiased predicate prediction caused by the long-tailed distribution. The prior works focus mainly on alleviating the deteriorating performances of the minority predicate predictions, showing drastic dropping recall scores, i.e., losing the majority predicate performances. It has not yet correctly analyzed the trade-off between majority and minority predicate performances in the limited SGG datasets. In this paper, to alleviate the issue, the Skew Class-balanced Re-weighting (SCR) loss function is considered for the unbiased SGG models. Leveraged by the skewness of biased predicate predictions, the SCR estimates the target predicate weight coefficient and then re-weights more to the biased predicates for better trading-off between the majority predicates and the minority ones. Extensive experiments conducted on the standard Visual Genome dataset and Open Image V4 \& V6 show the performances and generality of the SCR with the traditional SGG models.