Learning Condition-Invariant Scene Representations for Place Recognition across the Seasons Using Auto-Encoder and ICA

<div>“SUMMER” indicates that the training images are taken from the summer season. “{SUMMER, WINTER}” suggests that the training dataset contains images from the summer and winter seasons. Similar interpretations apply for the other two training datasets.</div>

Journal of Electrical and Computer Engineering

fig18

Figure 18

Figure 18: Learning Condition-Invariant Scene Representations for Place Recognition across the Seasons Using Auto-Encoder and ICA 