ai.djl

Interface Application.CV

Enclosing class:

Application

public static interface Application.CV

The common set of applications for computer vision (image and video data).

Field Summary

Fields

Modifier and Type	Field and Description
static Application	ACTION_RECOGNITION An application that accepts an image or video and classifies the action being done in it.
static Application	ANY Any computer vision application, including those in Application.CV.
static Application	IMAGE_CLASSIFICATION An application where images are assigned a single class name.
static Application	IMAGE_ENHANCEMENT An application that accepts an image and returns enhanced images.
static Application	IMAGE_GENERATION An application that accepts a seed and returns generated images.
static Application	INSTANCE_SEGMENTATION An application that finds zero or more objects in an image, the object class (see image classification), and their location as a pixel map.
static Application	OBJECT_DETECTION An application that finds zero or more objects in an image, the object class (see image classification), and their locations as a BoundingBox.
static Application	POSE_ESTIMATION An application that accepts an image of a single person and returns the Joints locations of the person.
static Application	SEMANTIC_SEGMENTATION An application that classifies each pixel in an image into a category.
static Application	WORD_RECOGNITION An application that accepts an image of a single word and returns the String text of the word.

Field Detail

ANY

static final Application ANY

Any computer vision application, including those in Application.CV.

IMAGE_CLASSIFICATION

static final Application IMAGE_CLASSIFICATION

An application where images are assigned a single class name.

Each image is given one of a fixed number of classes (or a probability of having that one class). The typical signature is Model<Image, Classifications>.

OBJECT_DETECTION

static final Application OBJECT_DETECTION

An application that finds zero or more objects in an image, the object class (see image classification), and their locations as a BoundingBox.

The typical signature is Model<Image, DetectedObjects>.

See Also:

The D2L chapter on object detection

SEMANTIC_SEGMENTATION

static final Application SEMANTIC_SEGMENTATION

An application that classifies each pixel in an image into a category.

INSTANCE_SEGMENTATION

static final Application INSTANCE_SEGMENTATION

An application that finds zero or more objects in an image, the object class (see image classification), and their location as a pixel map.

POSE_ESTIMATION

static final Application POSE_ESTIMATION

An application that accepts an image of a single person and returns the Joints locations of the person.

This can often be used with OBJECT_DETECTION to identify the people in the image and then run pose estimation on each detected person. The typical signature is Model<Image, Joints>.

ACTION_RECOGNITION

static final Application ACTION_RECOGNITION

An application that accepts an image or video and classifies the action being done in it.

WORD_RECOGNITION

static final Application WORD_RECOGNITION

An application that accepts an image of a single word and returns the String text of the word.

The typical signature is Model<Image, String>.

IMAGE_GENERATION

static final Application IMAGE_GENERATION

An application that accepts a seed and returns generated images.

The typical model returns an array of images Image[].

IMAGE_ENHANCEMENT

static final Application IMAGE_ENHANCEMENT

An application that accepts an image and returns enhanced images.

The typical signature is Model<Image, Image>.