Visual Question Answering Apparatus Using Hierarchical Visual Feature and Method Thereof

摘要

According to embodiments of the present invention, provided are a visual query-response apparatus and a method thereof. The visual query-response apparatus can improve not only the object recognition but also a correct answer rate with respect to the number or colors of objects, by inferring the correct answer of a visual question using a visual feature of a layer present at the lower level together with a feature of the final pooling layer among hierarchical features.