Quality-Aware Image-Text Alignment for Opinion-Unaware Image Quality Assessment
This addresses the scalability and applicability limitations of NR-IQA methods that rely on human annotations, offering a domain-specific solution for image quality assessment.
The paper tackles the problem of No-Reference Image Quality Assessment (NR-IQA) without needing human annotations by proposing QualiCLIP, a self-supervised method that aligns images with quality-related text prompts to generate quality-aware representations, achieving improved performance over existing opinion-unaware approaches and competitive results against supervised methods in cross-dataset experiments.
No-Reference Image Quality Assessment (NR-IQA) focuses on designing methods to measure image quality in alignment with human perception when a high-quality reference image is unavailable. Most state-of-the-art NR-IQA approaches are opinion-aware, i.e. they require human annotations for training. This dependency limits their scalability and broad applicability. To overcome this limitation, we propose QualiCLIP (Quality-aware CLIP), a CLIP-based self-supervised opinion-unaware approach that does not require human opinions. In particular, we introduce a quality-aware image-text alignment strategy to make CLIP generate quality-aware image representations. Starting from pristine images, we synthetically degrade them with increasing levels of intensity. Then, we train CLIP to rank these degraded images based on their similarity to quality-related antonym text prompts. At the same time, we force CLIP to generate consistent representations for images with similar content and the same level of degradation. Our experiments show that the proposed method improves over existing opinion-unaware approaches across multiple datasets with diverse distortion types. Moreover, despite not requiring human annotations, QualiCLIP achieves excellent performance against supervised opinion-aware methods in cross-dataset experiments, thus demonstrating remarkable generalization capabilities. The code and the model are publicly available at https://github.com/miccunifi/QualiCLIP.