RGB-D Individual Segmentation
This work addresses the challenge of fine-grained segmentation at the individual level, which is important for real-world applications like robotics and object recognition, but it appears incremental as it builds on existing segmentation methods with a new pipeline.
The paper tackles the problem of individual-level segmentation, where each object is a unique category with limited training data and unknown backgrounds, by proposing a Context Less-Aware (CoLA) pipeline that uses RGB-D data to reduce background context and incorporate 3D information. The results show that CoLA largely outperforms baseline methods on the YCB-Video and Supermarket-10K datasets.
Fine-grained recognition task deals with sub-category classification problem, which is important for real-world applications. In this work, we are particularly interested in the segmentation task on the \emph{finest-grained} level, which is specifically named "individual segmentation". In other words, the individual-level category has no sub-category under it. Segmentation problem in the individual level reveals some new properties, limited training data for single individual object, unknown background, and difficulty for the use of depth. To address these new problems, we propose a "Context Less-Aware" (CoLA) pipeline, which produces RGB-D object-predominated images that have less background context, and enables a scale-aware training and testing with 3D information. Extensive experiments show that the proposed CoLA strategy largely outperforms baseline methods on YCB-Video dataset and our proposed Supermarket-10K dataset. Code, trained model and new dataset will be published with this paper.