Intelligent image cutting method based on lightweight portrait detection

Abstract

The invention relates to an intelligent image cutting method based on lightweight portrait detection. The method comprises the following steps: S1, constructing a lightweight portrait detection network based on an anchor-free first-stage detection idea; s2, carrying out portrait detection on an input image according to the constructed lightweight portrait detection network, and obtaining position information of a portrait; and S3, according to the position information of the portrait, constraining the overall composition of the image by adopting a photographic composition principle, meanwhile, determining boundary information of the portrait, retaining complete information of a key figure, and realizing rapid cutting of the portrait picture without laying anchors. According to the automatic cutting algorithm designed according to the characteristics of the portrait picture, the difference between the portrait picture and the scenery picture is considered, and portrait detection and the photographic principle are combined, so that the portrait picture can be quickly and effectively processed, and the complete information of the key portrait can be reserved.

Description

technical field [0001] The invention relates to the field of automatic image cropping, in particular to an image intelligent cropping method based on lightweight portrait detection. Background technique [0002] Smart image cropping algorithms crop an image, given the viewport size or aspect ratio, so that the cropped result fits the viewport or aspect ratio, while ensuring that the most relevant or interesting parts are within the viewport. It can improve the visual quality of images and has a wide range of applications in many fields. The intelligent cropping algorithm can provide professional suggestions for image editing, saving a lot of image post-processing time; it can also help individuals quickly and accurately identify important content in images, and has an important application in the generation of image thumbnails. Directly shrink the entire image to fit the viewport, and smart cropping can generate a more ideal thumbnail for the user. [0003] The current alg...

AI Technical Summary

Problems solved by technology

Portraits occupy a large proportion in daily photography, and are the main body of users recording their daily life through social media. If the information of key figures is missing due to improper cropping methods, it will have a great impact on the final cropping effect, which is usually difficult for users to accept.

In order to solve the problem of lack of portrait information and slow cutting speed in existing automatic cropping methods, it is necessary to invent a new method for cropping portrait images to ensure that the cropping results are not only beautiful, but also retain key information, and the cropping speed is fast

Images