GSR-MAR: Global Super-Resolution for Person Multi-Attribute Recognition

Real-world single image super-resolution: A brief review

“…3) IFC [136]: The information fidelity criterion (IFC) [136] is a full-reference metric that assesses the quality of images based on natural scene statistics. 10 Research shows that the statistics of the space formed by natural images can be characterized using various models (e.g., the Gaussian scale mixtures). Generally, distortions would disturb the statistics of natural scenes and make images unnatural.…”

“…Compared to the hardware upgrade-based "hard" solution, the signal processing-based "soft" image resolution enhancement known as super-resolution (SR) is more flexible and economical. With the SR techniques that reconstruct a higher resolution output from the LR observation, we can obtain images with the resolution beyond the limit of imaging systems, thereby benefiting the subsequent analysis and understanding tasks such as segmentation [1]- [4], detection [5]- [7], and recognition [8]- [10].…”

Robust Pedestrian Attribute Recognition Using Group Sparsity for Occlusion Videos

“…Studies have been conducted on image-based PAR using various methods [2,3,6,16]. Liu et al [2] proposed the HydraPlus-Net network that utilizes multiscale features.…”

“…Zhao et al [3] proposed a recurrent convolutional architecture that understands the relationships between each group for PAR. Siadari et al [6] Feature Extractor…”

“…It is a human-searchable semantic description and can be used in soft biometrics for visual surveillance [1]. There have been several studies on this subject [2,3,4,5,6,7,8,9] because of the importance of its applications, such as finding missing persons and criminals. However, the occlusion problem is still under-handled.…”