Image and Video Caption Generator

Computer Science and Engineering

Images are visual memories that have a profound impact on the human encephalon, allowing us to recall specific information about a location, a certain person, or an object that we have captured in an instant. However, not all of the images are recognised, and a precise description of them is required to comprehend exactly what the image consists of. Deep learning and computer vision are used to understand the context of an image and classify it appropriately. It comprises labelling a photograph with English keywords by utilising datasets made available during model training. The CNN model Xception is trained using the imagenet dataset. Xception is in charge of image feature extraction. These recovered features will be used to modify the LSTM model in order to generate the image caption. Applications that automatically try to provide captions or explanations relating to photo and video frames have a lot of potential when using deep learning-based approaches. Captioning for photographs and videos is regarded as a key issue in imaging science. Among the application domains are general-purpose robot vision systems, automatically creating captions (or explanations) for images and videos for people with varying degrees of visual impairment, and many other application sectors. Each of these application groups would benefit greatly from additional task-concrete applications.

This work is licensed under a Creative Commons Attribution-ShareAlike 4.0 International License.