Aesthetic Attributes Assessment of Images

  title={Aesthetic Attributes Assessment of Images},
  author={Xin Jin and Le Wu and Geng Zhao and Xiaodong Li and Xiaokun Zhang and Shiming Ge and Dongqing Zou and Bin Zhou and Xinghui Zhou},
  journal={Proceedings of the 27th ACM International Conference on Multimedia},
  • Xin Jin, Le Wu, Xinghui Zhou
  • Published 11 July 2019
  • Computer Science
  • Proceedings of the 27th ACM International Conference on Multimedia
Image aesthetic quality assessment has been a relatively hot topic during the last decade. [] Key Method We introduce a new dataset named DPC-Captions which contains comments of up to 5 aesthetic attributes of one image through knowledge transfer from a full-annotated small-scale dataset. Then, we propose Aesthetic Multi-Attribute Network (AMAN), which is trained on a mixture of fully-annotated small-scale PCCD dataset and weakly-annotated large-scale DPC-Captions dataset.

Figures and Tables from this paper

Context-aware Attention Network for Predicting Image Aesthetic Subjectivity
This work explores a context-aware attention module in two dimensions: hierarchical and spatial that predicts the distribution of human aesthetic ratings of images, which reflects the diversity and similarity of human subjective opinions.
Understanding Aesthetics with Language: A Photo Critique Dataset for Aesthetic Assessment
The Reddit Photo Critique Dataset (RPCD), which contains tuples of image and photo critiques, is proposed, which is the first attempt to estimate the aesthetic quality of visual stimuli from the critiques and exploits the polarity of the sentiment of criticism as an indicator of aesthetic judgment.
A Comprehensive Survey on Computational Aesthetic Evaluation of Visual Art Images: Metrics and Challenges
This survey presents a comprehensive listing of the reviewed works on aesthetic assessment of photographs and paintings, mainly highlighting the contributions and innovations of the existing approaches and systematically evaluating recent deep learning techniques that are useful for developing robust models for aesthetic prediction tasks.
To be an Artist: Automatic Generation on Food Image Aesthetic Captioning
This paper proposes a novel model to generate aesthetic captions for food image aesthetic captioning as a compositional task that consists of two separated modules, i.e., a single-aspect captioning and an unsupervised text compression.
Distilling Knowledge from Object Classification to Aesthetics Assessment
This work proposes to distill knowledge on semantic patterns for a vast variety of image contents from multiple pre-trained object classification (POC) models to an IAA model to easier learn to relate various distinct contents to a limited number of aesthetic labels.
Image Composition Assessment with Saliency-augmented Multi-pattern Pooling
A composition assessment network SAMP-Net is proposed with a novel Saliency-Augmented Multi-pattern Pooling (SAMP) module, which analyses visual layout from the perspectives of multiple composition patterns, and extends Earth Mover’s Distance loss to weighted EMD loss to eliminate the content bias.
Multimodal Image Aesthetic Prediction with Missing Modality
A large number of experiments on two large benchmark datasets in the field of image aesthetic quality evaluation demonstrate that the proposed model significantly improves image aesthetic assessment performance under both textual missing modality condition and full-modality condition.
Aesthetic Evaluation and Guidance for Mobile Photography
This is the first system that can teach mobile users to shoot good photos in the form of aesthetic dashboard, through which, users can adjust several aesthetic attributes to take good photos easily.
Photo Stand-Out: Photography with Virtual Character
A novel optimization framework is proposed to synthesize an aesthetic pose for the virtual character with respect to the presented user's pose by optimizing a cost function that guides the rotation of each body joint angles.
Making Images Real Again: A Comprehensive Survey on Deep Image Composition
A comprehensive survey over the sub-tasks of image composition, which summarizes the traditional methods, deep learning based methods, datasets and evaluation, and points out the limitations of existing methods in each sub-task and the problem of the whole image composition task.


Aesthetic Critiques Generation for Photos
This work extends the image captioning task to produce captions related to photo aesthetics and/or photography skills and introduces a new dataset for aesthetics captioning called the Photo Critique Captioning Dataset (PCCD), which contains pair-wise image-comment data from professional photographers.
Neural Aesthetic Image Reviewer
This work proposes a model, referred to as Neural Aesthetic Image Reviewer, which can not only give an aesthetic score for an image, but also generate a textual description explaining why the image leads to a plausible rating score.
Distribution-Oriented Aesthetics Assessment With Semantic-Aware Hybrid Network
This paper proposes quantifying image aesthetics by a distribution over multiple quality levels, and presents a novel semantic-aware hybrid NEtwork (SANE), which harvests the information from object categorization and scene recognition to enhance image aesthetics assessment.
Deep Aesthetic Quality Assessment With Semantic Information
This paper addresses the correlation issue between automatic aesthetic quality assessment and semantic recognition, and develops an effective strategy to keep a balance between the two tasks, which facilitates to optimize the parameters of the framework.
Hierarchical aesthetic quality assessment using deep convolutional neural networks
A multi-scene deep learning model for image aesthetic evaluation
Photo Aesthetics Ranking Network with Attributes and Content Adaptation
This work proposes to learn a deep convolutional neural network to rank photo aesthetics in which the relative ranking of photo aesthetics are directly modeled in the loss function.
Joint Image and Text Representation for Aesthetics Analysis
Experimental results indicate that the proposed joint representation improves the performance of aesthetics assessment on the benchmarking AVA dataset, comparing with only visual features.
Deep image aesthetics classification using inception modules and fine-tuning connected layer
A novel DCNN structure codenamed ILGNet is proposed for image aesthetics classification, which introduces the Inception module and connects intermediate Local layers to the Global layer for the output and experimental results show that the proposed ILG net outperforms the state of the art results in image aesthetics assessment in the AVA benchmark.
Predicting Aesthetic Score Distribution through Cumulative Jensen-Shannon Divergence
A novel CNN based on the Cumulative distribution with Jensen-Shannon divergence (CJS-CNN) is presented to predict the aesthetic score distribution of human ratings, with a new reliability-sensitive learning methodbased on the kurtosis of the score distribution, which eliminates the requirement of the original full data of human rating data (without normalization).