Boosting Salient Object Detection with Transformer-based Asymmetric Bilateral U-Net

变压器 计算机科学 编码器 人工智能 卷积神经网络 突出 模式识别(心理学) 计算机视觉 工程类 电压 电气工程 操作系统
作者
Yu Qiu,Yun Liu,Le Zhang,Haotian Lu,Jing Xu
出处
期刊:IEEE Transactions on Circuits and Systems for Video Technology [Institute of Electrical and Electronics Engineers]
卷期号:: 1-1 被引量:13
标识
DOI:10.1109/tcsvt.2023.3307693
摘要

Existing salient object detection (SOD) methods mainly rely on U-shaped convolution neural networks (CNNs) with skip connections to combine the global contexts and local spatial details that are crucial for locating salient objects and refining object details, respectively. Despite great successes, the ability of CNNs in learning global contexts is limited. Recently, the vision transformer has achieved revolutionary progress in computer vision owing to its powerful modeling of global dependencies. However, directly applying the transformer to SOD is suboptimal because the transformer lacks the ability to learn local spatial representations. To this end, this paper explores the combination of transformers and CNNs to learn both global and local representations for SOD. We propose a transformer-based Asymmetric Bilateral U-Net (ABiU-Net). The asymmetric bilateral encoder has a transformer path and a lightweight CNN path, where the two paths communicate at each encoder stage to learn complementary global contexts and local spatial details, respectively. The asymmetric bilateral decoder also consists of two paths to process features from the transformer and CNN encoder paths, with communication at each decoder stage for decoding coarse salient object locations and fine-grained object details, respectively. Such communication between the two encoder/decoder paths enables AbiU-Net to learn complementary global and local representations, taking advantage of the natural merits of transformers and CNNs, respectively. Hence, ABiU-Net provides a new perspective for transformer-based SOD. Extensive experiments demonstrate that ABiU-Net performs favorably against previous state-of-the-art SOD methods. The code is available at https://github.com/yuqiuyuqiu/ABiU-Net.
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
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
VERY发布了新的文献求助10
1秒前
mym发布了新的文献求助10
2秒前
ringo完成签到,获得积分10
6秒前
7秒前
8秒前
VERY完成签到,获得积分20
9秒前
10秒前
小小王发布了新的文献求助10
12秒前
ZHONK1NG发布了新的文献求助10
13秒前
完美世界应助焦杨波采纳,获得10
13秒前
郭念坤发布了新的文献求助10
14秒前
14秒前
Akim应助moya采纳,获得10
14秒前
科研小白发布了新的文献求助10
14秒前
15秒前
17秒前
快乐滑板应助科研通管家采纳,获得10
19秒前
快乐滑板应助科研通管家采纳,获得10
19秒前
丘比特应助科研通管家采纳,获得10
19秒前
19秒前
快乐滑板应助科研通管家采纳,获得100
19秒前
pioneer应助科研通管家采纳,获得50
19秒前
SYLH应助科研通管家采纳,获得10
19秒前
快乐滑板应助科研通管家采纳,获得10
19秒前
科研通AI2S应助科研通管家采纳,获得10
19秒前
完美世界应助慵懒的猫采纳,获得10
20秒前
快乐滑板应助科研通管家采纳,获得10
20秒前
ceeray23应助科研通管家采纳,获得10
20秒前
pioneer应助科研通管家采纳,获得50
20秒前
aaaaaa发布了新的文献求助10
20秒前
WSY发布了新的文献求助10
20秒前
21秒前
22秒前
无心的冰巧应助焦杨波采纳,获得30
23秒前
英姑应助wise111采纳,获得10
24秒前
25秒前
lulu2024完成签到,获得积分10
25秒前
shu发布了新的文献求助10
26秒前
Bazinga完成签到,获得积分20
26秒前
27秒前
高分求助中
Production Logging: Theoretical and Interpretive Elements 2500
Востребованный временем 2500
Hopemont Capacity Assessment Interview manual and scoring guide 1000
Classics in Total Synthesis IV: New Targets, Strategies, Methods 1000
Neuromuscular and Electrodiagnostic Medicine Board Review 700
中介效应和调节效应模型进阶 400
Refractive Index Metrology of Optical Polymers 400
热门求助领域 (近24小时)
化学 医学 材料科学 生物 工程类 有机化学 生物化学 纳米技术 内科学 物理 化学工程 计算机科学 复合材料 基因 遗传学 物理化学 催化作用 细胞生物学 免疫学 电极
热门帖子
关注 科研通微信公众号,转发送积分 3443563
求助须知:如何正确求助?哪些是违规求助? 3039816
关于积分的说明 8978149
捐赠科研通 2728242
什么是DOI,文献DOI怎么找? 1496456
科研通“疑难数据库(出版商)”最低求助积分说明 691594
邀请新用户注册赠送积分活动 689113