文献类型：期刊文献

英文题名：Long-Tail Instance Segmentation Based on Memory Bank and Confidence Calibration

作者：Fan, Xinyue[1,2];Liu, Teng[1,2];Bao, Hong[1,2];Pan, Weiguo[1,2];Liang, Tianjiao[1,2];Li, Han[1,2]

第一作者：Fan, Xinyue

通讯作者：Bao, H[1];Pan, WG[1];Bao, H[2];Pan, WG[2]

机构：[1]Beijing Union Univ, Beijing Key Lab Informat Serv Engn, Beijing 100101, Peoples R China;[2]Beijing Union Univ, Coll Robot, Beijing 100101, Peoples R China

第一机构：北京联合大学北京市信息服务工程重点实验室

通讯机构：[1]corresponding author), Beijing Union Univ, Beijing Key Lab Informat Serv Engn, Beijing 100101, Peoples R China;[2]corresponding author), Beijing Union Univ, Coll Robot, Beijing 100101, Peoples R China.|[1141739]北京联合大学机器人学院;[11417]北京联合大学;[11417103]北京联合大学北京市信息服务工程重点实验室;

年份：2022

卷号：12

期号：18

外文期刊名：APPLIED SCIENCES-BASEL

收录：;Scopus(收录号：2-s2.0-85138611073);WOS:【SCI-EXPANDED(收录号:WOS:000859436800001)】；

基金：This research was funded by the National Natural Science Foundation of China (Nos. 61802019, 61932012, 61871039, 61906017 and 62006020), the Academic Research Projects of Beijing Union University (No. ZK10202202), the Beijing Municipal Education Commission Science and Technology Program (Nos. KM201911417003, KM201911417009 and KM201911417001), the Beijing Union University Research and Innovation Projects for Postgraduates (No.YZ2020K001), and the Premium Funding Project for Academic Human Resources Development in Beijing Union University under Grant BPHR2020DZ02.

语种：英文

外文关键词：long-tail; instance segmentation; Mask R-CNN; memory bank

摘要：In the field of computer vision, training a well-performing model on a dataset with a long-tail distribution is a challenging task. To address this challenge, image resampling is usually introduced as a simple and effective solution. However, when performing instance segmentation tasks, there may be multiple classes in one image. Hence, image resampling alone is not enough to obtain a sufficiently balanced distribution at the level of target data volume. In this paper, we propose an improved instance segmentation method for long-tail datasets based on Mask R-CNN. Specifically, an object-centric memory bank is used to establish an object-centric storage strategy that can solve the imbalance problem with respect to categories. In the testing phase, a post-processing calibration is used to adjust each class logit to change the confidence score, which improves the prediction score of tail classes. A discrete cosine transform-based mask is used to obtain high-quality masks, which improves segmentation accuracy. The evaluation of the proposed method on the LUIS dataset demonstrates its effectiveness. The proposed method improves the AP performance of EQL by 2.2%.