DeepAI

Adds color to a black and white image and outputs the colorized version.

Arbitrary Style Transfer in Real-time using Tensorflow

Detects if an image is NSFW. Returns a number between 0 and 1, with 1 being 100% likely to be NSFW.

Exaggerates feature attributes or textures in images.

This algorithm classifies each sentence in the input on a scale from very negative to very positive.

Compares two images to detect how visually similar they are.

Extracts the most relevant and unique words from a sample of text. Useful for classification.

Reduces the size of a document by only keeping the most relevant sentences from it.

This is a much faster implementation of "Neural Style" by pre-training on specific style examples.

Combines Context-Encoder with style transfer to apply a high quality texture to a hole in an image.

Transfers the style from one image onto the content of another image.

Creates an image from a line sketch. Only supports images of shoes currently.

Enhances the resolution of an input image.

Parses an input sentence in English and outputs the parse tree.

Summarizes the content of an image in a one sentence description.

Transfers the style from one image onto the content of another image.

Fills in a a rectangular hole centered in an image to make the filled in image look natural.

Performs upscaling and denoising on an input image.

Generates a collection of image boundaries for object detection.

Creates an image from scratch from a text description.

Captions an image by labeling every object the model detects is present within the image.