Face super-resolution reconstruction method and system based on progressive difference complementation

The invention discloses a face super-resolution reconstruction method and system based on progressive difference complementation, and the method comprises the steps: collecting a to-be-detected monitoring video, and extracting a low-resolution face image of a pedestrian from the monitoring video; ca...

Ausführliche Beschreibung

Gespeichert in:
Bibliographische Detailangaben
Hauptverfasser: ZHANG YUJIAO, CHEN ZHENXUE, QIN HAO, ZHONG KUNRU, CAO JIAQIAN
Format: Patent
Sprache:chi ; eng
Schlagworte:
Online-Zugang:Volltext bestellen
Tags: Tag hinzufügen
Keine Tags, Fügen Sie den ersten Tag hinzu!
Beschreibung
Zusammenfassung:The invention discloses a face super-resolution reconstruction method and system based on progressive difference complementation, and the method comprises the steps: collecting a to-be-detected monitoring video, and extracting a low-resolution face image of a pedestrian from the monitoring video; carrying out preprocessing on the low-resolution human face image; the preprocessed low-resolution image is input into a progressive difference complementary face super-resolution reconstruction network, and after the low-resolution face image is input into the network, a high-resolution feature image group and a low-resolution feature image group are output through first-stage coarse processing; outputting a final high-resolution feature image by the high-resolution feature image group and the low-resolution feature image group through difference value complementation in the second stage; and carrying out super-resolution reconstruction on the output final high-resolution feature image, carrying out pixel-by-pixel a