User-Specific Visual Attention Estimation Based on Visual Similarity and Spatial Information in Images

Few-Shot Personalized Saliency Prediction using Person Similarity based on Collaborative Multi-Output Gaussian Process Regression

“…Sim↑ KLdiv↓ CC↑ Signature [1] 0.412 8.04 0.413 GBVS [2] 0.447 6.89 0.437 Itti [3] 0.391 9.04 0.322 SalGAN [4] 0.569 3.56 0.635 Baseline1 [29] 0.503 4.13 0.597 Baseline2 [30] 0.417 7.64 0.401 FPSP based on similarity [18] [23]. Among the images in the dataset, 500 images were randomly selected as test images, and the remaining 1100 images were used as training images.…”

“…Furthermore, we adopted the following three PSM prediction methods using small amount of gaze data. Baseline1: A PSM prediction method using relationships of parts of the image and the whole image [29] Baseline2: A PSM prediction method using visual similarities [30] PSM based on similarity: A PSM prediction method similar to the proposed method. The difference between this method and our method is that this method used simple similarity of gaze tendency based on correlation coefficient of PSMs between the target person and other persons [18] Note that these PSM prediction methods were trained with only selected images via AIS in the same way as our method.…”

Few-Shot Personalized Saliency Prediction with Similarity of Gaze Tendency Using Object-Based Structural Information

“…Methods Sim↑ KLdiv↓ CC↑ Signature [3] 0.412 8.04 0.413 GBVS [4] 0.447 6.89 0.437 Itti [5] 0.391 9.04 0.322 SalGAN [6] 0.569 3.56 0.635 Contextual [28] 0.580 3.57 0.674 Baseline1 [31] 0.503 4.13 0.597 Baseline2 [32] 0.417 7.64 0.401 Similarity-based FPSP [19] [25], and its settings of the mini-batch size, the momentum, the number of layers L, the learning rate, and iterations were 9, 0.9, 3, 3.0×10 −5 , and 1000, respectively. Furthermore, the calculation method of SU(X) was the average of the PSMs of the training persons to eliminate the effect of USM calculation errors.…”

“…In addition to the USM prediction methods, we used the following four PSM prediction methods with a small amount of eye tracking data. Baseline1: A PSM prediction method based on both local and global information of input images [31]. Baseline2: A PSM prediction method based on visual similarities of the target and training images [32].…”

Few-Shot Personalized Saliency Prediction Based on Adaptive Image Selection Considering Object and Visual Attention

“…Consequently, FPSP based on AIS for the new target person can be realized from a small amount of training data with high accuracy. It should be noted that this paper is an extended version of [22]. Specifically, we enable the novel PSM prediction of the target person from those predicted from similar persons based on the multi-task CNN.…”