An automatic photo selection and trimming for graduation albums

In creating graduation albums, designers manually select appropriate photos from vast collections and then edit them into visually appealing images through cropping. These tasks require significant time and effort. Previous research has issues, such as false detections in face blur detection, making it impractical for graduation album production. Therefore, in this study, for graduation album photos, we first determine the main subject group, then apply face blur detection and blink detection to the face region to select the subjects. Next, it automatically trims photos, removing occluded subjects and unnecessary areas.