MUSCLE WP 5 ETeam on Visual Saliency

1

• MUSCLE  WP 5  E-Team on Visual Saliency
•  Participant   ICCS-NTUA  
• Researchers
• P. Maragos, I. Kokkinos, K. Rapantzikos, A.
  Sofou  Research Directions
• Scale-space feature detection  with Applications
  to object representation detection.
• Spatio-temporal saliency detection in video
  streams.
• Salient feature extraction and region-growing
  segmentation.

2
Scale-space feature detection  with Applications
to Object representation detection - I.

• Extract edge ridge lines in a scale-invariant
  manner
• Use simple descriptors to represent the extracted
  curves

3
Scale-space feature detection  with Applications
to Object representation detection -II.

• Use the extracted descriptors as a concise
  representation of the image.
• Explore the potential of using line and ridge
  features as the input to an object detection
  system.
• Compare combine with blob-like features.
• Evaluate on object detection tasks
• Horses
• Faces
• Cars

4
Spatiotemporal Visual Attention I Video Analysis

• Create video volume
• Feature extraction from spatiotemporal data
• Fusion saliency generation

5
Spatiotemporal Visual Attention II
Classification segmentation

• Use spatiotemporal VA for efficient global
  classification of videos
• Claim features extracted only from low or high
  saliency regions are more representative of the
  input video
• Foreground/Background segmentation
• Claim most salient regions are related to
  foreground areas of the video

6
Salient feature extraction and region-growing
segmentation I
Non- linear salient feature space
Salient feature regions extraction and region
growing segmentation result
7
Salient Feature Extraction and
Region-growing Segmentation II

• Salient feature extraction using linear
  (Gaussian scale-spaces) and non-linear
  methodologies (morphological scale spaces, AMFM
  models, multiband Energy tracking demodulation)
• Region-growing segmentation using salient
  features as leading marker-seeds.