图1 图像美学评估是计算摄影体系中图像质量评估的关键一环
图像的视觉美学质量衡量了在人类眼中一幅图像的视觉吸引力。由于视觉美学是一个主观的属性,往往会涉及情感和个人品味,这使得自动评估图像美学质量是一项非常主观的任务。然而,人们往往会达成一种共识,即一些图像在视觉上比其他图像更有吸引力。图像美学评估(Image Aesthetics Assessment, IAA)探索如何用可计算技术来预测人类对视觉刺激产生的情绪反应,使计算机模仿人类的审美过程,从而用可计算方法来自动预测图像的美学质量。
实验室的图像美学评估(Image Aesthetics Assessment, IAA)的研究内容包括:
1、总体IAA;
2、基于非完美数据的IAA;
3、多美学因素、多模态的IAA;
4、可解释性的IAA;
5、知识、数据、模型、价值四轮驱动的IAA;
6、面向AIGC的IAA;
7、面向人像摄影的IAA;
8、面向生活场景的IAA/AA。
实验室部分研究成果:图像美学评估开源代码&数据集。
vRobotit实验室关于“图像美学评估”代表性论文:
[1] Shuai He, Shuntian Zheng, Anlong Ming*, Yanni Wang, Huadong Ma, DA3Attacker: A Diffusion-based Attacker against Aesthetics-oriented Black-box Models, IEEE Transactions on Image Processing (TIP), revision, 2025.
开源项目地址:整理中
[2] Haobin Zhong, Shuai He, Anlong Ming*, Huadong Ma, Rethinking Personalized Aesthetics Assessment: Employing Physique Aesthetics Assessment as An Exemplification, IEEE Conference on Computer Vision and Pattern Recognition (CVPR), 2025.
开源项目地址:整理中
[3] Shuai He, Shuntian Zheng, Anlong Ming*, Banyu Wu, Huadong Ma, Rethinking No-reference Image Exposure Assessment from Holism to Pixel: Models, Datasets and Benchmarks, in Proceedings of the 38th International Conference on Neural Information Processing Systems (NIPS), 2024.
开源项目地址:https://github.com/mRobotit/Pixel-level-No-reference-Image-Exposure-Assessment
[4] Rui Xie, Anlong Ming*, Shuai He, Yi Xiao, Huadong Ma, "Special Relativity" of Image Aesthetics Assessment: a Preliminary Empirical Perspective, in Proceedings of the 32th ACM International Conference on Multimedia (MM), 2024.
开源项目地址:https://github.com/mRobotit/SR-IAA-image-aesthetics-and-quality-assessment
[5] Limin Liu, Shuai He, Anlong Ming*, Rui Xie, Huadong Ma, ELTA: An Enhancer against Long-Tail for Aesthetics-oriented Models, in Proceedings of the 41th international Conference on Machine Learning (ICML), 2024.
开源项目地址:https://github.com/mRobotit/Long-Tail-image-aesthetics-and-quality-assessment/
[6] Shai He, Anlong Ming*, Shuntian Zheng, Haobin Zhong, Huadong Ma, EAT: An Enhancer for Aesthetics-Oriented Transformers, in Proceedings of the 31th ACM International Conference on Multimedia (MM), 2023.
开源项目地址: https://github.com/mRobotit/Image-Aesthetics-Assessment/
[7] Shai He, Anlong Ming*, Yaqi Li, Jinyuan Sun, ShunTian Zheng, Huadong Ma, Thinking Image Color Aesthetics Assessment: Models, Datasets and Benchmarks, in Proceedings of International Conference on Computer Vision (ICCV), 2023.
开源项目地址: https://github.com/mRobotit/Image-Color-Aesthetics-Assessment
[8] Shuai He, Yongchang Zhang, Dongxiang Jiang, Rui Xie, Anlong Ming*, Rethinking Image Aesthetics Assessment: Models, Datasets and Benchmarks,the 31st International Joint Conference on Artificial Intelligence and the 25th European Conference on Artificial Intelligence (IJCAI), 2022.
开源项目地址:https://github.com/mRobotit/TANet
[9] Shuai He, Yi Xiao, Anlong Ming*, Huadong Ma, Prompt-Guided Image Color Aesthetics Assessment: Models, Datasets and Benchmarks,submitted to Information Fusion, accepted for publication, 2024.
开源项目地址:https://github.com/mRobotit/DeT-Plus