Most blind image quality assessment(BIQA)methods require a large amount of time to collect human opinion scores as training labels,which limits their usability in practice.Thus,we present an opinion-unaware BIQA metho...Most blind image quality assessment(BIQA)methods require a large amount of time to collect human opinion scores as training labels,which limits their usability in practice.Thus,we present an opinion-unaware BIQA method based on deep reinforcement learning which is trained without subjective scores,named DRL-IQA.Inspired by the human visual perception process,our model is formulated as a quality reinforced agent,which consists of the dynamic distortion generation part and the quality perception part.By considering the image distortion degradation process as a sequential decision-making process,the dynamic distortion generation part can develop a strategy to add as many different distortions as possible to an image,which enriches the distortion space to alleviate overfitting.A reward function calculated from quality degradation after adding distortion is utilized to continuously optimize the strategy.Furthermore,the quality perception part can extract rich quality features from the quality degradation process without using subjective scores,and accurately predict the state values that represent the image quality.Experimental results reveal that our method achieves competitive quality prediction performance compared to other state-of-the-art BIQA methods.展开更多
Abstract Chinese air pollution has increased in this century along with the rapid socioeconomic development and resulting anthropogenic emissions. While recent emission control measures have shown encouraging re sults...Abstract Chinese air pollution has increased in this century along with the rapid socioeconomic development and resulting anthropogenic emissions. While recent emission control measures have shown encouraging re sults and have reduced the levels of sulfur dioxide and primary aerosols, the concentrations of other air pollutants continue to grow, particularly secondary pollutants in cluding ozone and secondary aerosols. Meanwhile, a va riety of intentional and unintentional socioeconomic events have temporarily changed the pace, and even the signs, of growth of air pollution. These events include the short-term emission restrictions imposed during the Sino-African Summit, the Beijing Olympics and Para lympics, the Shanghai World Exposition （Shanghai Expo）, the Guangzhou Asian Olympics, and the Shenzhen Uni versiade, as well as the unintentional emission reductions associated with the recent economic recession and the annual Chinese New Year. This paper presents a brief overview of trends and temporary perturbations of Chi nese air pollution since 2000, summarizing studies on anthropogenic emission inventories, atmospheric meas urements, and inverse modeling. It concludes with rec ommendations for future research.展开更多
基金supported by the Fundamental Research Funds for the Central Universities.
文摘Most blind image quality assessment(BIQA)methods require a large amount of time to collect human opinion scores as training labels,which limits their usability in practice.Thus,we present an opinion-unaware BIQA method based on deep reinforcement learning which is trained without subjective scores,named DRL-IQA.Inspired by the human visual perception process,our model is formulated as a quality reinforced agent,which consists of the dynamic distortion generation part and the quality perception part.By considering the image distortion degradation process as a sequential decision-making process,the dynamic distortion generation part can develop a strategy to add as many different distortions as possible to an image,which enriches the distortion space to alleviate overfitting.A reward function calculated from quality degradation after adding distortion is utilized to continuously optimize the strategy.Furthermore,the quality perception part can extract rich quality features from the quality degradation process without using subjective scores,and accurately predict the state values that represent the image quality.Experimental results reveal that our method achieves competitive quality prediction performance compared to other state-of-the-art BIQA methods.
基金supported by the National Natural Science Foundation of China (Grant Nos. 41005078 and 41175127)
文摘Abstract Chinese air pollution has increased in this century along with the rapid socioeconomic development and resulting anthropogenic emissions. While recent emission control measures have shown encouraging re sults and have reduced the levels of sulfur dioxide and primary aerosols, the concentrations of other air pollutants continue to grow, particularly secondary pollutants in cluding ozone and secondary aerosols. Meanwhile, a va riety of intentional and unintentional socioeconomic events have temporarily changed the pace, and even the signs, of growth of air pollution. These events include the short-term emission restrictions imposed during the Sino-African Summit, the Beijing Olympics and Para lympics, the Shanghai World Exposition （Shanghai Expo）, the Guangzhou Asian Olympics, and the Shenzhen Uni versiade, as well as the unintentional emission reductions associated with the recent economic recession and the annual Chinese New Year. This paper presents a brief overview of trends and temporary perturbations of Chi nese air pollution since 2000, summarizing studies on anthropogenic emission inventories, atmospheric meas urements, and inverse modeling. It concludes with rec ommendations for future research.
