Meta-Seg: A Generalized Meta-Learning Framework for Multi-Class Few-Shot Semantic Segmentation

Abstract

Semantic segmentation performs pixel-wise classification for given images, which can be widely used in autonomous driving, robotics, medical diagnostics and etc. The recent advanced approaches have witnessed rapid progress in semantic segmentation. However, these supervised learning based methods rely heavily on large-scale datasets to acquire strong generalizing ability, such that they are coupled with some constraints. Firstly, human annotation of pixel-level segmentation masks is laborious and timeconsuming, which causes relatively expensive training data and make it hard to deal with urgent tasks in dynamic environment. Secondly, the outstanding performance of the above data-hungry methods will decrease with few available training examples. In order to overcome the limitations of the supervised learning semantic segmentation methods, this paper proposes a generalized meta-learning framework, named Meta-Seg. It consists of a meta-learner and a base-learner. Specifically, the meta-learner learns a good initialization and a parameter update strategy from a distribution of few-shot semantic segmentation tasks. The baselearner can be any semantic segmentation models theoretically and can implement fast adaptation (that is updating parameters with few iterations) under the guidance of the meta-learner. In this work, the successful semantic segmentation model FCN8s is integrated into Meta-Seg. Experiments on the famous few-shot semantic segmentation dataset PASCAL5 i prove Meta-Seg is a promising framework for few-shot semantic segmentation. Besides, this method can provide with reference for the relevant researches of meta-learning semantic segmentation.

Enhancement of Target‐Oriented Opinion Words Extraction with Multiview‐Trained Machine Reading Comprehension Model

“…MAML [ 36 ] seeks to learn a model initialization trained on multiple tasks and could be adapted to the target task with a few annotated data points. It has been widely used in the field of computer vision, involving visual tracking [ 37 ], incremental object detection [ 38 ], and semantic segmentation [ 39 ]. MAML has also been applied to many NLP tasks such as text classification [ 40 ], named entity recognition [ 41 ], and relation extraction [ 42 ].…”

Air-Conditioning Load Forecasting for Prosumer Based on Meta Ensemble Learning

“…where F(x p ) represents the final prediction result;f i x p and C i x p denote the forecasting result of the i-th point forecaster and its weight. Considering the prediction model in Fig.1, we can obtain the weight parameters of meta ensemble learning by minimizing the energy function of the training samples [35]. In this paper, we define the energy function as the sum of the mean square error of the training samples and the weight decay term, as shown below:…”

Meta-DRN: Meta-Learning for 1-Shot Image Segmentation

“…Although meta-learning has seen a lot of success with few-shot image classification, frequently reaching stateof-the-art results, these techniques have not been as extensively used for segmentation. Some examples include [17], [18]. We use the FSS-1000 [7] dataset for our models.…”