# Visual Relationship Detection with Language prior and Softmax

## TL;DR

This paper introduces a visual relationship detection method that leverages language priors and a softmax classifier, achieving state-of-the-art results without complex linguistic knowledge distillation.

## Contribution

It proposes a novel approach combining language priors with visual features and a simple softmax classifier, avoiding complex loss functions or large text corpus distillation.

## Key findings

- Outperforms previous state-of-the-art methods on Visual Relationship Detection.
- Effective use of language priors improves relationship classification accuracy.
- Simplifies the model by avoiding complex loss functions and large-scale linguistic knowledge distillation.

## Abstract

Visual relationship detection is an intermediate image understanding task that detects two objects and classifies a predicate that explains the relationship between two objects in an image. The three components are linguistically and visually correlated (e.g. "wear" is related to "person" and "shirt", while "laptop" is related to "table" and "on") thus, the solution space is huge because there are many possible cases between them. Language and visual modules are exploited and a sophisticated spatial vector is proposed. The models in this work outperformed the state of arts without costly linguistic knowledge distillation from a large text corpus and building complex loss functions. All experiments were only evaluated on Visual Relationship Detection and Visual Genome dataset.