Stable Learning for Neural Network Tomography by Using Back Projected Image.

Masaru Teranishi Keita Oka Masahiro Aramoto

Published in: DCAI (2014)

Keyphrases

neural network
image data
multiscale
input image
single image
learning algorithm
reinforcement learning
image content
image pixels
pixel values
template matching
image regions
image representation
linear predictors
image collections
active learning
high resolution
prior knowledge
image analysis
image segmentation
supervised learning
online learning
object detection
image features
keypoints
spatial information
learning process
incremental learning
image retrieval