Method and apparatus for conversion into sketches for geometrical representation learning

An apparatus for conversion into sketches for geometrical representation learning, the apparatus comprising:

• an input/output interface configured to receive an image and output results of processing of the image;
• a storage configured to store a program for performing a method for conversion into sketches;
• and
• a controller including at least one processor, and the controller configured to convert the image into a sketch by executing the program,
• wherein the storage stores an encoder, a stroke generation model comprising the encoder, and a stroke embedding model connected to the stroke generation model, wherein the stroke generation model includes a first decoder, a second decoder connected to the first decoder, and an intermediate layer connected to the first decoder,
• wherein the controller:
• receives initial stroke information having attribute parameters and image information about a target image;
• converts the target image into a plurality of final strokes having geometric information about the target image by using a stroke generation model;
• and
• outputs results of data processing of the plurality of final strokes as representations in a vector format,
• wherein the encoder extracts image features from the target image,
• wherein the first decoder receives strokes and the image features from the encoder and outputs converted strokes,
• wherein the intermediate layer receives the converted strokes from the first decoder and extracts intermediate strokes,
• wherein the second decoder receives the converted strokes from the first decoder and outputs the plurality of final strokes,
• wherein the stroke embedding model outputs a representation of all strokes obtained by mapping the plurality of final strokes received from the second decoder to an embedding space, and
• wherein, in a process of outputting the representations in the vector format, the controller is configured to:
• convert the image features output by the encoder into a first representation in a vector format;
• output a second representation in a vector format by combining parameters of each stroke included in the plurality of final strokes into a vector format;
• output a third representation in a vector format regarding the representation of all strokes output by the stroke embedding model;
• select the first representation, the second representation, and the third representation;
• and
• combine the first representation, the second representation, and the third representation.