Transformer-Based Approach Via Contrastive Learning for Zero-Shot Detection

计算机科学 人工智能 变压器 模式识别(心理学) 零(语言学) 工程类 电气工程 语言学 哲学 电压
作者
Wei Liu,Hui Chen,Yongqiang Ma,Jianji Wang,Nanning Zheng
出处
期刊:International Journal of Neural Systems [World Scientific]
卷期号:33 (07) 被引量:5
标识
DOI:10.1142/s0129065723500351
摘要

Zero-shot detection (ZSD) aims to locate and classify unseen objects in pictures or videos by semantic auxiliary information without additional training examples. Most of the existing ZSD methods are based on two-stage models, which achieve the detection of unseen classes by aligning object region proposals with semantic embeddings. However, these methods have several limitations, including poor region proposals for unseen classes, lack of consideration of semantic representations of unseen classes or their inter-class correlations, and domain bias towards seen classes, which can degrade overall performance. To address these issues, the Trans-ZSD framework is proposed, which is a transformer-based multi-scale contextual detection framework that explicitly exploits inter-class correlations between seen and unseen classes and optimizes feature distribution to learn discriminative features. Trans-ZSD is a single-stage approach that skips proposal generation and performs detection directly, allowing the encoding of long-term dependencies at multiple scales to learn contextual features while requiring fewer inductive biases. Trans-ZSD also introduces a foreground-background separation branch to alleviate the confusion of unseen classes and backgrounds, contrastive learning to learn inter-class uniqueness and reduce misclassification between similar classes, and explicit inter-class commonality learning to facilitate generalization between related classes. Trans-ZSD addresses the domain bias problem in end-to-end generalized zero-shot detection (GZSD) models by using balance loss to maximize response consistency between seen and unseen predictions, ensuring that the model does not bias towards seen classes. The Trans-ZSD framework is evaluated on the PASCAL VOC and MS COCO datasets, demonstrating significant improvements over existing ZSD models.
最长约 10秒,即可获得该文献文件

科研通智能强力驱动
Strongly Powered by AbleSci AI
更新
PDF的下载单位、IP信息已删除 (2025-6-4)

科研通是完全免费的文献互助平台,具备全网最快的应助速度,最高的求助完成率。 对每一个文献求助,科研通都将尽心尽力,给求助人一个满意的交代。
实时播报
笨笨友桃完成签到,获得积分10
刚刚
在水一方应助Huang采纳,获得10
刚刚
1秒前
2秒前
郭郭要努力ya完成签到 ,获得积分10
3秒前
4秒前
科研通AI2S应助Saint采纳,获得10
4秒前
annafan完成签到,获得积分10
4秒前
5秒前
科目三应助雷桑采纳,获得10
5秒前
少吃顿饭并不难完成签到 ,获得积分10
6秒前
珠珠发布了新的文献求助10
6秒前
6秒前
刘三哥完成签到 ,获得积分10
7秒前
隐形曼青应助leederay采纳,获得10
8秒前
上官若男应助to_ooooo采纳,获得10
8秒前
海阔天空发布了新的文献求助10
9秒前
顺心羊完成签到,获得积分10
9秒前
科研_小白完成签到,获得积分10
9秒前
XJTU_jyh完成签到,获得积分10
10秒前
TheaGao完成签到 ,获得积分10
10秒前
不是省油的灯完成签到 ,获得积分10
12秒前
14秒前
14秒前
bububu完成签到,获得积分10
15秒前
nanaki完成签到,获得积分10
15秒前
16秒前
小蘑菇应助lin采纳,获得10
17秒前
17秒前
18秒前
czcz发布了新的文献求助10
18秒前
Sean完成签到 ,获得积分10
21秒前
雷桑发布了新的文献求助10
21秒前
窗户上的喵咪很无聊完成签到 ,获得积分10
22秒前
Saint完成签到,获得积分10
23秒前
23秒前
零点起步完成签到,获得积分10
24秒前
认真丹亦完成签到 ,获得积分10
25秒前
时光完成签到,获得积分10
26秒前
27秒前
高分求助中
【提示信息,请勿应助】关于scihub 10000
Les Mantodea de Guyane: Insecta, Polyneoptera [The Mantids of French Guiana] 3000
徐淮辽南地区新元古代叠层石及生物地层 3000
The Mother of All Tableaux: Order, Equivalence, and Geometry in the Large-scale Structure of Optimality Theory 3000
Global Eyelash Assessment scale (GEA) 1000
Picture Books with Same-sex Parented Families: Unintentional Censorship 550
Research on Disturbance Rejection Control Algorithm for Aerial Operation Robots 500
热门求助领域 (近24小时)
化学 材料科学 医学 生物 工程类 有机化学 生物化学 物理 内科学 纳米技术 计算机科学 化学工程 复合材料 遗传学 基因 物理化学 催化作用 冶金 细胞生物学 免疫学
热门帖子
关注 科研通微信公众号,转发送积分 4038619
求助须知:如何正确求助?哪些是违规求助? 3576294
关于积分的说明 11375058
捐赠科研通 3306084
什么是DOI,文献DOI怎么找? 1819374
邀请新用户注册赠送积分活动 892698
科研通“疑难数据库(出版商)”最低求助积分说明 815066