Unsupervised place discovery for visual place classification
| Autor: | Hao Guoqing, Tanaka Kanji, Fei Xiaoxiao, Inamoto Kouya |
|---|---|
| Rok vydání: | 2017 |
| Předmět: |
0209 industrial biotechnology
Recall Computer science business.industry Pattern recognition 02 engineering and technology Workspace 010501 environmental sciences 01 natural sciences Convolutional neural network Partition (database) 020901 industrial engineering & automation Robot Artificial intelligence business Classifier (UML) Robotic mapping 0105 earth and related environmental sciences |
| Zdroj: | MVA |
| DOI: | 10.23919/mva.2017.7986802 |
| Popis: | In this study, we explore the use of deep convolutional neural network (DCNN) in visual place classification for robotic mapping and localization. An open question is how to partition the robot's workspace into places so as to maximize the performance (e.g., accuracy, precision & recall) of potential DCNN classifiers. This is a chicken and egg problem: If we had a well-trained DCNN classifier, it is rather easy to partition the robot's workspace into places, but the training of a DCNN classifier requires a set of pre-defined place classes. In this study, we address this problem and present several strategies for unsupervised discovery of place classes (“time cue”, “location cue”, “time-appearance cue”, and “location-appearance cue”) and evaluate efficacy of the proposed methods using publicly available NCLT dataset. |
| Databáze: | OpenAIRE |
| Externí odkaz: |
|