A method for automatic cropping of pictures

Abstract

The invention discloses a method for automatically cutting pictures. The specific steps are: calculating the picture ratio by using the height and width information of the picture to be processed; setting the output size of the picture to be processed, and determining the output ratio of the picture to be processed according to the output size; Determine the size of the cropping frame with the output ratio; judge whether the image ratio of the image to be processed is appropriate to the required output ratio; predict the category of the image to be processed and detect the salient area of ​​the image to be processed, and extract the salient feature map of the image to be processed; Distinguish the more important area and the less important area according to the salient feature map, and obtain the minimum circumscribed rectangle of the more important area; judge whether the minimum circumscribed rectangle is covered by the cropping box, perform cropping or blank processing, and output the cropping result. The invention has the advantages of automatically cutting pictures, being applicable to the output of cutting results of any size, judging whether pictures should be cut or left blank, and performing blank processing on pictures that cannot be cut.

Description

technical field [0001] The invention relates to the technical fields of deep learning and computer vision, in particular to a method for automatically cutting pictures. Background technique [0002] In the field of image printing, when image printing is performed, when the aspect ratio of the image is different from the aspect ratio of the actual image, it is necessary to preprocess the image. The preprocessing principles are as follows: first try to crop the image according to the printing size, and judge whether the cropped image maintains the integrity of the characters, landmarks, special characters, and composition of the original image. Integrity can be cropped, otherwise it needs to be left blank. Blank means that the original image is not cropped, and white pixels are filled in the vacant positions of the photo paper during printing. All pictures are first tried to be cropped. If the integrity of the picture is destroyed after cropping, it can be left blank. [000...

AI Technical Summary

Problems solved by technology

[0005] The above three methods are only applicable to its own application field, but not applicable to the field of picture printing, because the content of the picture to be printed is complex and diverse, and the picture containing the face may only contain part of the face or the head has hair ornaments, hats etc.; there may be text, date, human gestures, landmark buildings, etc. in the picture. These areas are not allowed to be cropped when the picture is printed, and the cropping position cannot be correctly determined by face detection or background recognition alone; the second And the third method has no requirements on the size of the final cropping frame, which can be adjusted arbitrarily according to the detection results, which does not meet the requirements of image cropping in the field of image printing. about the size of

Images