Generation and Comprehension of Unambiguous Object Descriptions
This work addresses the challenge of ambiguous object descriptions for computer vision and natural language processing applications, representing an incremental improvement over existing methods.
The authors tackled the problem of generating and comprehending unambiguous object descriptions in images, showing that their method outperforms previous approaches that ignore scene ambiguity, with results based on a new large-scale dataset derived from MS-COCO.
We propose a method that can generate an unambiguous description (known as a referring expression) of a specific object or region in an image, and which can also comprehend or interpret such an expression to infer which object is being described. We show that our method outperforms previous methods that generate descriptions of objects without taking into account other potentially ambiguous objects in the scene. Our model is inspired by recent successes of deep learning methods for image captioning, but while image captioning is difficult to evaluate, our task allows for easy objective evaluation. We also present a new large-scale dataset for referring expressions, based on MS-COCO. We have released the dataset and a toolbox for visualization and evaluation, see https://github.com/mjhucla/Google_Refexp_toolbox