Label or Message: A Large-Scale Experimental Survey of Texts and Objects Co-Occurrence
This work provides foundational data for computer vision applications involving text-object relationships, but it is incremental as it applies existing methods to new data without introducing novel techniques.
The paper conducted a large-scale survey of co-occurrence between visual objects and scene texts using a large image dataset and state-of-the-art scene text detection/recognition tools, focusing on 'label' texts attached to objects to detail them. The analysis revealed statistics about label texts and insights into how scene texts aid object recognition and vice versa.
Our daily life is surrounded by textual information. Nowadays, the automatic collection of textual information becomes possible owing to the drastic improvement of scene text detectors and recognizer. The purpose of this paper is to conduct a large-scale survey of co-occurrence between visual objects (such as book and car) and scene texts with a large image dataset and a state-of-the-art scene text detector and recognizer. Especially, we focus on the function of "label" texts, which are attached to objects for detailing the objects. By analyzing co-occurrence between objects and scene texts, it is possible to observe the statistics about the label texts and understand how the scene texts will be useful for recognizing the objects and vice versa.