DeepScores - A Dataset for Segmentation, Detection and Classification of Tiny Objects.
Lukas TuggenerIsmail EleziJürgen SchmidhuberMarcello PelilloThilo StadelmannPublished in: CoRR (2018)
Keyphrases
- feature extraction and classification
- foreground background separation
- cluttered scenes
- object segmentation
- multiple objects
- bounding box
- classification accuracy
- features extraction
- benchmark datasets
- complex background
- image segmentation
- level set
- support vector machine
- multiscale
- support vector
- feature set
- image classification
- segmentation algorithm
- outdoor images
- detection method
- feature selection
- category labels
- detecting objects
- object detection
- training dataset
- decision trees
- figure ground segmentation
- page segmentation
- moving objects
- pixel classification
- training data
- cluttered background
- object detectors
- training set
- false positives
- outdoor scenes
- d objects
- feature extraction
- feature space
- image regions
- detection algorithm
- text classification