Method of image completion
A method of image completion comprises: constructing the image repair model and constructing a plurality of conditional generative adversarial networks according to a plurality of object types; inputting the training image corresponding to the plurality of objective types such that the plurality of conditional generative adversarial networks respectively conduct a corruption feature training; inputting the image in need of repair and respectively conducting an image repair through the plurality of conditional generative adversarial networks to generate a plurality of repaired images; and judging a reasonable probability of the plurality of repaired images through a probability analyzer, choosing an accomplished image and outputting the accomplished image through an output interface.
1. A method of image completion, comprising:
configuring, in an image processor, an image repair model and constructing a plurality of conditional generative adversarial networks corresponding to a plurality of object types;
inputting respectively, by an input interface, training images corresponding to the plurality of object types such that each of the plurality of conditional generative adversarial networks conducts corruption feature training respectively;
inputting, by the input interface, an image in need of repair and conducting image repair through the plurality of conditional generative adversarial networks to generate a plurality of repaired images respectively; and
computing, by a first judging model, a reasonable probability of each of the plurality of repaired images, choosing an accomplished image and outputting the accomplished image through an output interface.
2. The method of image completion of claim 1 , wherein the image repair model comprises an image transferring model and a second judging model; the image transferring model transfers the image in need of repair to a sample image; and the second judging model judges whether or not the sample image conforms to a real image.
3. The method of image completion of claim 1 , wherein the corruption feature training further comprises steps below:
generating a plurality of corruption types by a mask processor;
conducting destruction of the plurality of corruption types to the training image to form a plurality of corrupted images; and
amending the image repair model through the plurality of corrupted images and the originally inputted training image.
4. The method of image completion of claim 3 , wherein the plurality of corruption types comprise text corruption, line corruption, scribble corruption, random corruption or corruption of arbitrary polygons.
5. The method of image completion of claim 3 , wherein the destruction to the training image comprises rotating the plurality of corrupted images with a predetermined angle.
6. The method of image completion of claim 1 , wherein the step of inputting the image in need of repair through the input interface further comprises steps below:
marking, by the input interface, a repair area on the image in need of repair; repairing the repair area of the image in need of repair through the plurality of conditional generative adversarial networks; and generating the plurality of repaired images.
7. The method of image completion of claim 1 , wherein the step of inputting the image in need of repair through the input interface further comprises steps below:
marking, by the input interface, an erasure area on the image in need of repair; erasing an image of the erasure area and then conducting the image repair to the erasure area through the plurality of conditional generative adversarial networks; and generating the plurality of repaired images.
8. The method of image completion of claim 1 , wherein the plurality of object types comprise a car, a human being, a dog, a tree, a road or combination thereof.
9. The method of image completion of claim 1 , wherein the plurality of object types comprise facial features of human being, accessories or combination thereof.
10. The method of image completion of claim 1 , wherein the plurality of conditional generative adversarial networks are stored in a database respectively; the image processor connects to each of the databases and simultaneously accesses the plurality of conditional generative adversarial networks for conducting the image repair.