TAMIL LANGUAGE IMAGE ANNOTATION USING SEMANTIC ORDER LEARNING AND ATTENTION BASED NEURAL MACHINE TRANSLATION

Nowadays, picture and sentence matching has made tremendous strides, but because of the current broad visual- semantic gap, it remains difficult. The words in the phrases are structured in a grammatical way sequentially, while the semantic meanings in the images are typically unorganised. Generally, the image and phrase matching task relates to calculating the visual-semantic resemblance between an image and a phrase. Here, the words in the phrases are structured in a grammatical way sequentially, while the semantic meanings in the images are typically unorganised. Thus, semantic concepts are suggested and an order learning system for the matching of images and phrases in this work, which can enhance the representation of images by first predicting semantic concepts and then arranging them in the correct semantic order. Further the sentence generated is translated to Tamil language since there is no image annotation system for the same.