Image Caption via Visual Attention Switch on DenseNet

Image caption enables computers to generate a text description of images automatically. However, the generated description is not good enough recently. Computers can describe what objects are in the image but cannot give more details about these objects. In this study, we present a novel image caption approach to give more details when describing objects. In detail, a visual attention-based LSTM is used to find the objects, as well as a semantic attention-based LSTM is used for giving semantic attributes. At last, a gated object-attribute matching network is used to match the objects to their semantic attributes. The experiments on the public datasets of Flickr30k and MSCOCO demonstrate that the proposed approach improved the quality of the image caption, compared with the most advanced methods at present.

Download Full-text

What Grasps and Holds 8-Month-Old Infants' Looking Attention? The Effects of Object Size and Depth Cues

Child Development Research ◽

10.1155/2012/439618 ◽

2012 ◽

Vol 2012 ◽

pp. 1-10 ◽

Cited By ~ 4

Author(s):

Yu Guan ◽

Daniela Corbetta

Keyword(s):

Visual Attention ◽

Eye Tracking ◽

Sustained Attention ◽

Object Size ◽

Relative Impact ◽

Depth Cues ◽

The Times ◽

Attention Switch

The current eye-tracking study explored the relative impact of object size and depth cues on 8-month-old infants' visual attention processes. A series of slides containing 3 objects of either different or same size were displayed on backgrounds with varying depth cues. The distribution of infants' first looks (a measure of initial attention switch) and infants' looking durations (a measure of sustained attention) at the objects were analyzed. Results revealed that the large objects captured infants' attention first, that is, most of the times infants directed their visual attention first to the largest object in the scene regardless of depth cues. For sustained attention, infants preferred maintaining their attention to the largest object also, but this occurred only when depth cues were present. These findings suggest that infants' initial attention response is driven mainly by object size, while infants' sustained attention is more the product of combined figure and background processing, where object sizes are perceived as a function of depth cues.

Download Full-text

Visual Spatial Attention in Migraine Sufferers in Postictal and Interictal Phases: An Event-Related Potential Study

Journal of Psychophysiology ◽

10.1027//0269-8803.15.1.22 ◽

2001 ◽

Vol 15 (1) ◽

pp. 22-34 ◽

Cited By ~ 2

Author(s):

D.H. de Koning ◽

J.C. Woestenburg ◽

M. Elton

Keyword(s):

Cerebral Cortex ◽

Visual Attention ◽

Migraine Attack ◽

Spatial Attention ◽

Event Related Potential ◽

Significant Enhancement ◽

Pathophysiological Mechanism ◽

Attention Task ◽

Visual Spatial Attention ◽

Visual Spatial

Migraineurs with and without aura (MWAs and MWOAs) as well as controls were measured twice with an interval of 7 days. The first session of recordings and tests for migraineurs was held about 7 hours after a migraine attack. We hypothesized that electrophysiological changes in the posterior cerebral cortex related to visual spatial attention are influenced by the level of arousal in migraineurs with aura, and that this varies over the course of time. ERPs related to the active visual attention task manifested significant differences between controls and both types of migraine sufferers for the N200, suggesting a common pathophysiological mechanism for migraineurs. Furthermore, migraineurs without aura (MWOAs) showed a significant enhancement for the N200 at the second session, indicating the relevance of time of measurement within migraine studies. Finally, migraineurs with aura (MWAs) showed significantly enhanced P240 and P300 components at central and parietal cortical sites compared to MWOAs and controls, which seemed to be maintained over both sessions and could be indicative of increased noradrenergic activity in MWAs.

Download Full-text

Visual Attention: Fitting It All Together

Contemporary Psychology ◽

10.1037/000289 ◽

1997 ◽

Vol 42 (6) ◽

pp. 501-503

Author(s):

Kyle R. Cave

Keyword(s):

Visual Attention

Download Full-text