Transformer Lesion Tracker
نویسندگان
چکیده
Evaluating lesion progression and treatment response via longitudinal tracking plays a critical role in clinical practice. Automated approaches for this task are motivated by prohibitive labor costs time consumption when matching is done manually. Previous methods typically lack the integration of local global information. In work, we propose transformer-based approach, termed Transformer Lesion Tracker (TLT). Specifically, design Cross Attention-based (CAT) to capture combine both information enhance feature extraction. We also develop Registration-based Anatomical Attention Module (RAAM) introduce anatomical CAT so that it can focus on useful knowledge. A Sparse Selection Strategy (SSS) presented selecting features reducing memory footprint training. addition, use regression further improve model performance. conduct experiments public dataset show superiority our method find performance has improved average Euclidean center error at least 14.3% (6mm vs. 7mm) compared with state-of-the-art (SOTA). Code available https://github.com/TangWen920812/TLT.
منابع مشابه
GLAST Tracker ∗
The Large Area Telescope (LAT) on board the Gamma-ray Large-Area Space Telescope (GLAST) is a pair-conversion gamma-ray detector designed to explore the gamma-ray universe in the 20 MeV– 300 GeV energy band. The Tracker subsystem of the LAT will perform tracking of electron and positrons to determine the origin of the gamma-ray. The design and performance of the GLAST LAT Tracker are described ...
متن کاملVIS-Tracker: A Wearable Vision-Inertial Self-Tracker
We present a demonstrated and commercially viable self-tracker, using robust software that fuses data from inertial and vision sensors. Compared to infrastructurebased trackers, self-trackers have the advantage that objects can be tracked over an extremely wide area, without the prohibitive cost of an extensive network of sensors or emitters to track them. So far most AR research has focused on...
متن کاملTransformer proteins
Proteins are generally believed to adopt a unique fold, defined by their amino acid sequence, under specific environmental conditions. These unique structures, in turn, endow proteins with one specific function. However, not all proteins obey the “1 amino acid sequence → 1 fold → 1 function” scheme. Moonlighting proteins that adopt one distinct threedimensional structure but can accomplish two ...
متن کاملImage Transformer
Image generation has been successfully cast as an autoregressive sequence generation or transformation problem. Recent work has shown that self-attention is an effective way of modeling textual sequences. In this work, we generalize a recently proposed model architecture based on self-attention, the Transformer, to a sequence modeling formulation of image generation with a tractable likelihood....
متن کاملذخیره در منابع من
با ذخیره ی این منبع در منابع من، دسترسی به آن را برای استفاده های بعدی آسان تر کنید
ژورنال
عنوان ژورنال: Lecture Notes in Computer Science
سال: 2022
ISSN: ['1611-3349', '0302-9743']
DOI: https://doi.org/10.1007/978-3-031-16446-0_19