Color is an essential and stable feature of vehicles. It serves as a reliable cue in a wide range of applications of intelligent transportation systems. In this paper, we present deep learning models for the recognition of color and vehicles. The vehicle color captioning model introduces a wide variety of color names mimicking human-level color name annotations. Unlike the existing color recognition models, which tag color names from a limited set of pre-defined color names, the proposed color recognition model employs a generative model to give color captions to vehicles. The generative model is a sequence-to-sequence model that considers color naming as a machine translation problem. The proposed model for color naming does not need a labelled ground-truth dataset. Rather, it generates color captions from mean values of the pixels and adopts color name adjectives like ‘dark’, ‘light’, ‘dull’, etc. Along with color recognition, we also present a pre-trained model vehicle recognition, which recognizes the vehicle in transportation.
Deep Generative Model for Vehicle Color Recognition
2024-12-17
1742750 byte
Conference paper
Electronic Resource
English
Vehicle Color Recognition With Spatial Pyramid Deep Learning
Online Contents | 2015
|A novel statistical generative model dedicated to face recognition
British Library Online Contents | 2010
|