Abstract
Since the state-of-the-art deep learning algorithms demand a large training dataset, which is often unavailable in some domains, the transfer of knowledge from one domain to another has been a trending technique in the computer vision field. However, this method may not be a straight-forward task considering several issues such as original network size or large differences between the source and target domain. In this paper, we perform transfer learning for semantic segmentation of off-road driving environments using a pre-trained segmentation network called DeconvNet. We explore and verify two important aspects regarding transfer learning. First, since the original network size was very large and did not perform well for our application, we proposed a smaller network, which we call the light-weight network. This light-weight network is half the size to the original DeconvNet architecture. We transferred the knowledge from the pre-trained DeconvNet to our light-weight network and fine-tuned it. Second, we used synthetic datasets as the intermediate domain before training with the real-world off-road driving data. Fine-tuning the model trained with the synthetic dataset that simulates the off-road driving environment provides more accurate results for the segmentation of real-world off-road driving environments than transfer learning without using a synthetic dataset does, as long as the synthetic dataset is generated considering real-world variations. We also explore the issue whereby the use of a too simple and/or too random synthetic dataset results in negative transfer. We consider the Freiburg Forest dataset as a real-world off-road driving dataset.
Highlights
Semantic segmentation, a task based on pixel-level image classification, is a fundamental approach in the field of computer vision for scene understanding
As we are interested in the off-road autonomous driving scenario, we focused on how the transfer learning works in order to segment the real-world dataset with and without using synthetic dataset
We explored the transfer learning from the perspective of network size and training techniques with and without the use of synthetic data
Summary
A task based on pixel-level image classification, is a fundamental approach in the field of computer vision for scene understanding. As the objects of interest, such as roads, trees, and terrains, are continuous rather than discrete structures, detection algorithms often cannot give detailed information, hindering the performance of autonomous vehicles This is not true of semantic segmentation algorithms, as all the objects of interests are detected on a pixel-by-pixel basis. The convolutional layer is followed by a sub-sampling layer called the pooling layer This layer is supposed to capture the high-level information of feature maps in compressed form. It helps to make the features invariant to smaller transitions and translations which results in CNNs being capable of focusing on the useful properties and ignoring the less important features in the feature space. Max-pooling is the famous pooling technique which takes the maximum value of pixels within a defined boundary as its output
Sign-in/Register to view highlights & summary Sign-in/Register to access full text options 
Talk to us
Join us for a 30 min session where you can share your feedback and ask us any queries you have
Schedule a callDisclaimer: All third-party content on this website/platform is and will remain the property of their respective owners and is provided on "as is" basis without any warranties, express or implied. Use of third-party content does not indicate any affiliation, sponsorship with or endorsement by them. Any references to third-party content is to identify the corresponding services and shall be considered fair use under The CopyrightLaw.
