Scene Understanding

Scene Understanding

Combination of recent advances in CV and Machine Translation to produce image captions. Uses CNN, RNN and Transfer Learning.

Artificial Intelligence

  • 0 Collaborators

  • 0 Followers

Description

A model to directly generate the sequence of the words most relevant to the image conditioned on the image and previously generated words. Therefore, they can produce novel combinations of words that might never have occurred in the training data. Implementation of Google Show and Tell Model using Tensorflow.

Gallery

Links

Presentation on the Model Implemented and the Results obtained.

Medium 0 1jv2ntkuohgyo5bpcwwm audirccafbpxuem3xduwlc xiks3yemtrlfrs u4ihjkywfrzrfhxgbg9npttx  skuxxgcg9vj3txa9tu2o5pogtxk39qfrfyreo

Parag J. (Intel) created project Scene Understanding

Medium 0eb202c6 527e 4bb7 84ba a76d0bd4a686

Scene Understanding

A model to directly generate the sequence of the words most relevant to the image conditioned on the image and previously generated words. Therefore, they can produce novel combinations of words that might never have occurred in the training data. Implementation of Google Show and Tell Model using Tensorflow.

No users to show at the moment.

No users to show at the moment.