Bengali Visual Genome

Bengali Visual Genome is a multimodal dataset consisting of text and images suitable for English-to-Bengali multimodal machine translation task and multimodal research. We have selected short English segments (captions) from Visual Genome along with associated images and manually translated them to Bengali by Bengali native speaker, taking the associated images into account.

The training set contains 29K segments. Further 1K and 1.6K segments are provided in a development and test sets, respectively, which follow the same (random) sampling from the original Hindi Visual Genome. Additionally, a challenge test set of 1400 segments was prepared for the WAT2022 multi-modal task. This challenge test set was created by searching for (particularly) ambiguous English words based on the embedding similarity and manually selecting those where the image helps to resolve the ambiguity. The surrounding words in the sentence however also often include sufficient cues to identify the correct meaning of the ambiguous word.

Dataset Sample

Sample items from the randomly selected segments (train/dtest/etest)
Image	Image_ID	X	Y	Width	Height	English Text	Bengali Text
	2323457	20	150	325	121	Many giraffes at a zoo	একটি চিড়িয়াখানায় অনেক জিরাফ
	2335684	61	191	437	182	Fruit stand outside market	বাজারের বাইরে ফলের স্ট্যান্ড
Sample item from the challenge test set (chtest)
	2372733	26	107	152	218	The tennis court is made up of sand and dirt	টেনিস কোর্ট বালু এবং ময়লা দিয়ে গঠিত

Sample ambiguous word: court (the key different meanings are court of justice vs. tennis court). As the image example illustrates, the word alone is ambiguous but already the English caption may contain sufficient information to disambiguate the word; the image is thus not always necessary for correct translation even in our challenge test set.

Download Link

   http://hdl.handle.net/11234/1-3722