Improving the performance of multimedia pedestrian classification with images synthesized using a deep convolutional generative adversarial network.
Chuho YiJungwon ChoPublished in: Multim. Tools Appl. (2021)
Keyphrases
- image classification
- multimedia
- input image
- classification accuracy
- image data
- network architecture
- image database
- image features
- edge detection
- machine learning
- synthesized images
- pattern recognition
- image analysis
- unsupervised learning
- object recognition
- discriminative methods
- preprocessing stage
- content based retrieval
- deep learning
- image regions
- segmentation algorithm
- computer vision
- feature selection
- decision trees
- skin lesion
- image collections
- neural network
- feature vectors
- labelled training data
- test images
- generative model
- feature points
- multi view
- e learning
- support vector
- image registration
- semi supervised
- image retrieval
- high resolution