Traditional saliency reloaded: A good old model in new shape

Abstract

In this paper, we show that the seminal, biologicallyinspired saliency model by Itti et al. [21] is still competitive with current state-of-the-art methods for salient object segmentation if some important adaptions are made. We show which changes are necessary to achieve high performance, with special emphasis on the scale-space: we introduce a twin pyramid for computing Difference-of-Gaussians, which enables a flexible center-surround ratio. The resulting system, called VOCUS2, is elegant and coherent in structure, fast, and computes saliency at the pixel level. It is not only suitable for images with few objects, but also for complex scenes as captured by mobile devices. Furthermore, we integrate the saliency system into an object proposal generation framework to obtain segment-based saliency maps and boost the results for salient object segmentation. We show that our system achieves state-of-theart performance on a large collection of benchmark data.

BioVision: A Biomimetics Platform for Intrinsically Motivated Visual Saliency Learning

“…Saliency maps can be either purely bottom-up [61], [17], [25], or refined by top-down modulation [23], [62], [18], [20]. Bottom-up saliency highlights stimuli that are intrinsically salient in their context, which may sometimes be sufficient for scene exploration [64].…”

“…However, top-down modulation, which highlights elements that are relevant for a specific task, is more meaningful for the problem of object detection in indoor environments. Saliency maps are either fixation-based [30], [17] or area-based [11], [20], [61]. Fixation-based approach is related with the probability of a human being to make a fixation at a given pixel position, while area-based approach consider salient elements (typically objects) as a whole area of the image.…”

S-CNN-Based Ship Detection From High-Resolution Remote Sensing Images

“…The famous ITTI salient detection model (Itti and Koch, 2000) integrates gray feature, direction and color information to construct basic physiological visual similar features. However, the results with the orientation channel usually turns out to be less useful for salient object segmentation (Frintrop et al, 2015) since it assigns high saliency values to object edges, and makes object segmentation difficult. Thus we extract the same saliency features: color and intensity ones with (Frintrop et al, 2015) to generate the salient map.…”

“…However, the results with the orientation channel usually turns out to be less useful for salient object segmentation (Frintrop et al, 2015) since it assigns high saliency values to object edges, and makes object segmentation difficult. Thus we extract the same saliency features: color and intensity ones with (Frintrop et al, 2015) to generate the salient map. We extract color features with the intensity and newly defined RG and BY color channels, which are calculated as I = (R + G + B)/3, RG = R − G and BY = B − (R + G)/2, respectively.…”

Interpretable Feature Maps for Robot Attention

“…The input is represented as feature maps consisting of colour channels and responses of oriented Gabor filters. Recently, this approach has been modified to detect larger salient regions instead of points by Frintrop et al [5], showing the continued appeal of the approach. Other modifications include weighting the different feature maps after identifying useful features [11] and exploring the role of saliency in overt attention [19].…”