000221552 001__ 221552
000221552 005__ 20190317000537.0
000221552 020__ $$a978-1-4673-9961-6
000221552 0247_ $$2doi$$a10.1109/ICIP.2016.7532767
000221552 022__ $$a1522-4880
000221552 02470 $$2ISI$$a000390782002071
000221552 037__ $$aCONF
000221552 245__ $$aIMAGE AESTHETIC PREDICTORS BASED ON WEIGHTED CNNS
000221552 269__ $$a2016
000221552 260__ $$bIeee$$c2016$$aNew York
000221552 300__ $$a5
000221552 336__ $$aConference Papers
000221552 490__ $$aIEEE International Conference on Image Processing ICIP
000221552 520__ $$aConvolutional Neural Networks (CNNs) have been widely adopted for many imaging applications. For image aesthetics prediction, state-of-the-art algorithms train CNNs on a recently-published large-scale dataset, AVA. However, the distribution of the aesthetic scores on this dataset is extremely unbalanced, which limits the prediction capability of existing methods. We overcome such limitation by using weighted CNNs. We train a regression model that improves the prediction accuracy of the aesthetic scores over state-of-the-art algorithms. In addition, we propose a novel histogram prediction model that not only predicts the aesthetic score, but also estimates the difficulty of performing aesthetics assessment for an input image. We further show an image enhancement application where we obtain an aesthetically pleasing crop of an input image using our regression model.
000221552 6531_ $$aAesthetics
000221552 6531_ $$asample weights
000221552 6531_ $$aCNN
000221552 700__ $$0247134$$g221036$$aJin, Bin
000221552 700__ $$aOrtiz Segovia, Maria V.
000221552 700__ $$0241946$$g125681$$aSüsstrunk, Sabine
000221552 7112_ $$dSeptember 25-28$$cPhoenix, AZ, USA$$aThe 23rd IEEE International Conference on Image Processing (ICIP 2016)
000221552 773__ $$t2016 Ieee International Conference On Image Processing (Icip)$$q2291-2295
000221552 8564_ $$uhttps://infoscience.epfl.ch/record/221552/files/ICIP%202106-Jin.pdf$$zn/a$$s2401725$$yn/a
000221552 909C0 $$xU10429$$0252320$$pIVRL
000221552 909CO $$qGLOBAL_SET$$pconf$$ooai:infoscience.tind.io:221552$$pIC
000221552 917Z8 $$x114979
000221552 917Z8 $$x125681
000221552 937__ $$aEPFL-CONF-221552
000221552 973__ $$rREVIEWED$$sPUBLISHED$$aEPFL
000221552 980__ $$aCONF