The Design of High-Level Features for Photo Quality Assessment

By Yan Ke, Xiaoou Tang, Feng Jing. Computer Vision and Pattern Recognition, 2006

760 citations recorded by the source corpus.

Subfields: Computer Vision and Pattern Recognition

View Paper: https://www.semanticscholar.org/paper/6989f866a445bb880ed8663a05dc4cd71c87b1b7

How researchers see this paper

The researchers whose trusted networks rank it highest, strongest advocate first. Every researcher gets a different order.

In Computer Vision and Pattern Recognition

Citations in the corpus

Creating Summaries from User Videos (2014, methodology): “To predict the aesthetic quality of a frame we compute colorfulness [3], contrast [14] and the distribution of edges [14].”
Content-based photo quality assessment (2011, methodology): “Our complexity features outperform previous complexity features such as “simplicity”[ 12] and “hue count”[11].”
High level describable attributes for predicting aesthetics and interestingness (2011, methodology): “Describability is key here so that we can ask people to label images according to the presence or absence of an attribute and then use this labeled data to train classifiers for recognizing image attributes.”
Photo and Video Quality Evaluation: Focusing on the Subject (2008, methodology): “[6], and a large and diverse video database collected from professional movies and amateurish videos.”
A framework for photo-quality assessment and enhancement based on visual aesthetics (2010, background): “Most of the earlier papers [6, 7, 16, 12, 19, 18, 3, 17, 21] on this topic evaluate their respective approaches on their own private collections, which are not made available.”
Scene recognition and weakly supervised object localization with deformable part-based models (2011, methodology): “9% of its edge energy using a modifi cation of the technique from [9].”
Using Color Compatibility for Assessing Image Realism (2007, background): “[9] propose a set of high-level image features for assessing the perceived artistic quality of a photograph.”
RAPID: Rating Pictorial Aesthetics using Deep Learning (2014, background): “Texture descriptors vary from wavelet-based texture features [6], distribution of edges, to blur descriptors and shallow depth-of-ﬁeld descriptors [13].”
> Replace This Line with Your Paper Identification Number (double-click Here to Edit) < (background): “Features are extracted both globally and locally, while only global features based on every pixel are extracted in [1][3][5]; 3.”
AVA: A large-scale database for aesthetic visual analysis (2012, background): “Typically, features capturing the “aes-thetic” properties of an image are proposed with the aim of mimicking photographic rules and practices such as the golden ratio, the rule of thirds and color harmonies [3, 11, 16, 6, 15].”
The Interestingness of Images (2013, background): “These include image quality [15], memorability [14] and aesthetics [10, 11].”
Image Aesthetic Assessment: An experimental survey (2016, methodology): “[9] are among the first to cast aesthetic understanding of images into a binary classification problem.”
Interestingness Prediction by Robust Learning to Rank (2014, background): “Early efforts on image interestingness prediction focus on different aspects than interestingness as such, including image quality [22], memorability [19], and aesthetics [7].”
Assessing the aesthetic quality of photographs using generic image descriptors (2011, background): “Early works employed low level features inspired by perceptually motivated metrics [4] [12] [3].”
A Deep Network Solution for Attention and Aesthetics Aware Photo Cropping (2019, background): “Some works [44], [45], [46], [47], [48] use hand-crafted aesthetics”
Image retrieval: Ideas, influences, and trends of the new age (2008, result): “Another attempt [Ke et al. 2006] at distinguishing high-quality images from low-quality ones has found similar levels of success with data obtained from yet another peer-rated photo contest oriented Website [DPChallenge.”
Computing iconic summaries of general visual concepts (2008, methodology): “Since quality ranking is inherently subjective and existing methods [6, 14] are not perfectly reliable, it is difficult to say whether the top-ranked image selected by our method is in fact the “best” one to represent its cluster.”
Photo Aesthetics Ranking Network with Attributes and Content Adaptation (2016, methodology): “Datasets such as CUHKPQ [11,19] only provide binary labels (low or high aesthetics) which cannot easily be used for rating prediction.”