Paper of BIT Won Best Paper Award of EURASIP in 2020


  Recently, Associate Professor Liu Lixiong, Master Student Hua Yi, Professor Zhao Qingjie, Professor Huang Hua (corresponding author) of the School of Computer Science and Technology, Beijing Institute of Technology(BIT), and Professor Alan Conrad Bovik of the University of Texas at Austin "Adaboosting Neural Network" won the European Association for Signal Processing (EURASIP) 2020 Best Paper Award (2020 EURASIP Best Paper Award), and will be awarded at the 28th European Signal Processing Conference (EUSIPCO 2020) in January 2021.

Performance of OG-IQA algorithm on TID2013 image library

  The awarded paper was published in Signal Processing: Image Communication (SPIC), which is an important international  journal in this field, in January 2016, and was supported by the National Natural Science Foundation of China. This paper proposes a reference-free image quality evaluation method (OG-IQA) based on gradient correlation, which solves the problem that traditional image gradient information cannot accurately describe the degree of image distortion. This method expresses the traditional gradient information in the form of relative gradient changes, and uses the correlation between adjacent gradients to enhance the description of the degree of image distortion. Then, the AdaBoosting BP neural network is used to establish the mapping relationship between image features and image quality. This method can achieve great performance on multiple image libraries.

  The EURASIP SPIC journal chooses the best paper award every two years. The selection range is the papers published in the past four years (this time is from 2015 to 2018). The academic evaluation, download and citation of the papers are the main selection standard.