This work investigates three methods for calculating loss for autoencoder-based pretraining of image encoders: The commonly used reconstruction loss, the more recently introduced deep perceptual similarity loss, and a feature prediction loss proposed here; the latter turning out to be the most efficient choice. Standard auto-encoder pretraining for deep learning tasks is done by comparing the input image and the reconstructed image. Recent work shows that predictions based on embeddings generated by image autoencoders can be improved by training with perceptual loss, i.e., by adding a loss network after the decoding step. So far the autoencoders trained with loss networks implemented an explicit comparison of the original and reconstructed images using the loss network. However, given such a loss network we show that there is no need for the time-consuming task of decoding the entire image. Instead, we propose to decode the features of the loss network, hence the name "feature prediction loss". To evaluate this method we perform experiments on three standard publicly available datasets (LunarLander-v2, STL-10, and SVHN) and compare six different procedures for training image encoders (pixel-wise, perceptual similarity, and feature prediction losses; combined with two variations of image and feature encoding/decoding). The embedding-based prediction results show that encoders trained with feature prediction loss is as good or better than those trained with the other two losses. Additionally, the encoder is significantly faster to train using feature prediction loss in comparison to the other losses. The method implementation used in this work is available online: https://github.com/guspih/Perceptual-Autoencoders
University of Science and Technology of China（中国科学技术大学）；Microsoft Research Asia（微软亚洲研究院）
We propose bootstrapped masked autoencoders (BootMAE), a new approach for vision BERT pretraining. BootMAE improves the original masked autoencoders (MAE) with two core designs: 1) momentum encoder that provides online feature as extra BERT prediction targets; 2) target-aware decoder that tries to reduce the pressure on the encoder to memorize target-specific information in BERT pretraining. The first design is motivated by the observation that using a pretrained MAE to extract the features as the BERT prediction target for masked tokens can achieve better pretraining performance. Therefore, w…
Baidu VIS（百度视觉技术部）；Australian National University（澳大利亚国立大学）；Beihang University（北京航空航天大学）；Peking University（北京大学）
We present a strong object detector with encoder-decoder pretraining and finetuning. Our method, called Group DETR v2, is built upon a vision transformer encoder ViT-Huge~\cite{dosovitskiy2020image}, a DETR variant DINO~\cite{zhang2022dino}, and an efficient DETR training method Group DETR~\cite{chen2022group}. The training process consists of self-supervised pretraining and finetuning a ViT-Huge encoder on ImageNet-1K, pretraining the detector on Object365, and finally finetuning it on COCO. Group DETR v2 achieves $\textbf{64.5}$ mAP on COCO test-dev, and establishes a new SoTA on the COCO le…
JD AI Research（京东人工智能研究院）；Sun Yat-sen University（中山大学）
Despite having impressive vision-language (VL) pretraining with BERT-based encoder for VL understanding, the pretraining of a universal encoder-decoder for both VL understanding and generation remains challenging. The difficulty originates from the inherently different peculiarities of the two disciplines, e.g., VL understanding tasks capitalize on the unrestricted message passing across modalities, while generation tasks only employ visual-to-textual message passing. In this paper, we start with a two-stream decoupled design of encoder-decoder structure, in which two decoupled cross-modal enc…
King Abdullah University of Science and Technology (KAUST)（阿卜杜拉国王科技大学）
Due to the large memory footprint of untrimmed videos, current state-of-the-art video localization methods operate atop precomputed video clip features. These features are extracted from video encoders typically trained for trimmed action classification tasks, making such features not necessarily suitable for temporal localization. In this work, we propose a novel supervised pretraining paradigm for clip features that not only trains to classify activities but also considers background clips and global video information to improve temporal sensitivity. Extensive experiments show that using fea…
Semantic segmentation labels are expensive and time consuming to acquire. Hence, pretraining is commonly used to improve the label-efficiency of segmentation models. Typically, the encoder of a segmentation model is pretrained as a classifier and the decoder is randomly initialized. Here, we argue that random initialization of the decoder can be suboptimal, especially when few labeled examples are available. We propose a decoder pretraining approach based on denoising, which can be combined with supervised pretraining of the encoder. We find that decoder denoising pretraining on the ImageNet d…
Technical Report for the ICRA 2026 GOOSE 2D Fine-Grained Semantic Segmentation Challenge: Pretraining-Diverse Ensemble of Foundation Vision Encoders for Robust Outdoor Scene Understanding
LION Lab, Hefei University of Technology（合肥工业大学 LION 实验室）；University of Macau（澳门大学）
Sea AI Lab（Sea人工智能实验室）；National University of Singapore（新加坡国立大学）
For unsupervised pretraining, mask-reconstruction pretraining (MRP) approaches, e.g. MAE and data2vec, randomly mask input patches and then reconstruct the pixels or semantic features of these masked patches via an auto-encoder. Then for a downstream task, supervised fine-tuning the pretrained encoder remarkably surpasses the conventional ``supervised learning'' (SL) trained from scratch. However, it is still unclear 1) how MRP performs semantic feature learning in the pretraining phase and 2) why it helps in downstream tasks. To solve these problems, we first theoretically show that on an aut…
Wrong Organ, Right Physics: Transferring Echocardiography Pretraining to Lung Ultrasound for Tuberculosis Screening
University of Stellenbosch（斯泰伦博斯大学）；Swiss Federal Institute of Technology (EPFL)（瑞士联邦理工学院（EPFL））；National Teaching Hospital for Tuberculosis and Pulmonary Diseases (CNHU-PPC)（国家结核病与肺病教学医院（CNHU-PPC））；South African Medical Research Council（南非医学研究理事会）；Stellenbosch University（斯泰伦博斯大学）
The Chinese University of Hong Kong（香港中文大学）；NVIDIA AI Technology Center（英伟达人工智能技术中心）；SenseTime Research（商汤科技研究院）；Tsinghua University（清华大学）
FlowFormer introduces a transformer architecture into optical flow estimation and achieves state-of-the-art performance. The core component of FlowFormer is the transformer-based cost-volume encoder. Inspired by the recent success of masked autoencoding (MAE) pretraining in unleashing transformers' capacity of encoding visual representation, we propose Masked Cost Volume Autoencoding (MCVA) to enhance FlowFormer by pretraining the cost-volume encoder with a novel MAE scheme. Firstly, we introduce a block-sharing masking strategy to prevent masked information leakage, as the cost maps of neighb…
Purdue University（普渡大学）；Amazon AWS AI Labs（亚马逊AWS人工智能实验室）；OPPO US Research Center（OPPO美国研究中心）
Unsupervised learning of vision transformers seeks to pretrain an encoder via pretext tasks without labels. Among them is the Masked Image Modeling (MIM) aligned with pretraining of language transformers by predicting masked patches as a pretext task. A criterion in unsupervised pretraining is the pretext task needs to be sufficiently hard to prevent the transformer encoder from learning trivial low-level features not generalizable well to downstream tasks. For this purpose, we propose an Adversarial Positional Embedding (AdPE) approach -- It distorts the local visual structures by perturbing …
Tsinghua University（清华大学）；Beijing National Research Center for Information Science and Technology (BNRist)（北京信息科学与技术国家研究中心）；University of Maryland（马里兰大学）
Auto-encoder is an important architecture to understand point clouds in an encoding and decoding procedure of self reconstruction. Current auto-encoder mainly focuses on the learning of global structure by global shape reconstruction, while ignoring the learning of local structures. To resolve this issue, we propose Local-to-Global auto-encoder (L2G-AE) to simultaneously learn the local and global structure of point clouds by local to global reconstruction. Specifically, L2G-AE employs an encoder to encode the geometry information of multiple scales in a local region at the same time. In addit…
RadJEPA: Radiology Encoder for Chest X-Rays via Joint Embedding Predictive Architecture
Department of Computer Science and Engineering, Indian Institute of Technology Bombay（印度理工学院孟买分校计算机科学与工程系）；Department of Biomedical Engineering, Johns Hopkins University（约翰霍普金斯大学生物医学工程系）；Koita Centre for Digital Health, Indian Institute of Technology Bombay（印度理工学院孟买分校Koita数字健康中心）
LeVJEPA: Efficient & Scalable Video Pretraining without the Heuristics
German Cancer Research Center（德国癌症研究中心）；German Cancer Consortium（德国癌症联盟）；Goethe University Frankfurt（法兰克福大学）；Mila（米拉研究所）；Université de Montréal（蒙特利尔大学）；Brown University（布朗大学）；Courant Institute, New York University（纽约大学柯朗研究所）；Advanced Machine Intelligence (AMI Labs)（高级机器智能实验室（AMI Labs））
Reconstruction of Patient-Specific Confounders in AI-based Radiologic Image Interpretation using Generative Pretraining
Else Kroener Fresenius Center for Digital Health (EKFZ)（埃尔泽·克勒纳·菲森纽斯数字健康中心（EKFZ））；Technical University Dresden（德累斯顿工业大学）；University Hospital Aachen（亚琛大学医院）；UC Berkeley（加州大学伯克利分校）；University Hospital Zurich（苏黎世大学医院）；University of Zurich（苏黎世大学）；Center for Artificial Intelligence in Medicine and Imaging (AIMI)（医学与成像人工智能中心（AIMI））；Stanford University（斯坦福大学）；University Hospital Dresden（德累斯顿大学医院）；National Center for Tumor Diseases (NCT)（国家肿瘤疾病中心（NCT））
Efficient Self-supervised Vision Pretraining with Local Masked Reconstruction
King Abdullah University of Science and Technology（阿卜杜拉国王科技大学）；Monash University（莫纳什大学）；Nile University（尼罗河大学）；Nanyang Technological University（南洋理工大学）
PanoSplatt3R: Leveraging Perspective Pretraining for Generalized Unposed Wide-Baseline Panorama Reconstruction
School of Electronics and Information, Northwestern Polytechnical University and Shaanxi Key Laboratory of Information Acquisition and Processing（电子工程学院、西北工业大学和陕西省信息获取与处理重点实验室）
Efficient Reinforcement Learning Through Adaptively Pretrained Visual Encoder
Institute of Automation, Chinese Academy of Sciences（中国科学院自动化研究所）；University of Chinese Academy of Sciences（中国科学院大学）；Chinese Academy of Sciences（中国科学院）
Understanding Self-Supervised Pretraining with Part-Aware Representation Learning
Key Lab of High Confidence Software Technologies (Peking University), Ministry of Education, China（高可信软件技术教育部重点实验室（北京大学））；School of Computer Science, Peking University（北京大学计算机学院）；School of AI, Peking University（北京大学人工智能学院）；University of Hong Kong（香港大学）；School of EIC, Huazhong University of Science & Technology（华中科技大学电子信息与通信学院）；Baidu（百度）；UC Berkeley（加州大学伯克利分校）
Self-attention based transformer models have been dominating many computer vision tasks in the past few years. Their superb model qualities heavily depend on the excessively large labeled image datasets. In order to reduce the reliance on large labeled datasets, reconstruction based masked autoencoders are gaining popularity, which learn high quality transferable representations from unlabeled images. For the same purpose, recent weakly supervised image pretraining methods explore language supervision from text captions accompanying the images. In this work, we propose masked image pretraining…
Location-Aware Pretraining for Medical Difference Visual Question Answering
Department of Electrical and Computer Engineering, Carnegie Mellon University, Kigali, Rwanda（电气与计算机工程系，卡内基梅隆大学，刚果（金）基利齐）；University of Melbourne, Melbourne, Australia（墨尔本大学，墨尔本，澳大利亚）
ViTEraser: Harnessing the Power of Vision Transformers for Scene Text Removal with SegMIM Pretraining
South China University of Technology（华南理工大学）；SCUT-Zhuhai Institute of Modern Industrial Innovation（华南理工大学珠海现代产业创新研究院）；INTSIG-SCUT Joint Lab of Document Image Analysis and Recognition（INTSIG-华南理工大学文档图像分析与识别联合实验室）；Huazhong University of Science and Technology（华中科技大学）
Transmission Line Defect Detection Based on UAV Patrol Images and Vision-language Pretraining
Department of Electronic and Communication Engineering, North China Electric Power University（电子与通信工程系，华北电力大学）；Hebei Key Laboratory of Power Internet of Things Technology, North China Electric Power University（河北省电力物联网技术重点实验室，华北电力大学）；Computer Science and Software Engineering Department, Monmouth University（计算机科学与软件工程系，蒙特莫恩大学）
Huazhong University of Science and Technology（华中科技大学）
High angular resolution diffusion imaging (HARDI) demands a lager amount of data measurements compared to diffusion tensor imaging, restricting its use in practice. In this work, we explore a learning-based approach to reconstruct HARDI from a smaller number of measurements in q-space. The approach aims to directly learn the mapping relationship between the measured and HARDI signals from the collecting HARDI acquisitions of other subjects. Specifically, the mapping is represented as a 1D encoder-decoder convolutional neural network under the guidance of the compressed sensing (CS) theory for …
University College London（伦敦大学学院）；University of Bremen（不来梅大学）；University of Oulu（奥卢大学）；The Chinese University of Hong Kong（香港中文大学）
Deep image prior (DIP) was recently introduced as an effective unsupervised approach for image restoration tasks. DIP represents the image to be recovered as the output of a deep convolutional neural network, and learns the network's parameters such that the output matches the corrupted observation. Despite its impressive reconstructive properties, the approach is slow when compared to supervisedly learned, or traditional reconstruction techniques. To address the computational challenge, we bestow DIP with a two-stage learning paradigm: (i) perform a supervised pretraining of the network on a …
Northeastern University（东北大学）；NEC Labs America（NEC美国实验室）；UC San Diego（加州大学圣地亚哥分校）
Self-supervised vision-language pretraining from pure images and text with a contrastive loss is effective, but ignores fine-grained alignment due to a dual-stream architecture that aligns image and text representations only on a global level. Earlier, supervised, non-contrastive methods were capable of finer-grained alignment, but required dense annotations that were not scalable. We propose a single stream architecture that aligns images and language at multiple levels: global, fine-grained patch-token, and conceptual/semantic, using two novel tasks: symmetric cross-modality reconstruction (…
Thomas Jefferson National Accelerator Facility（托马斯·杰斐逊国家加速器实验室）
In this article we describe the development of machine learning models to assist the CLAS12 tracking algorithm by identifying tracks through inferring missing segments in the drift chambers. Auto encoders are used to reconstruct missing segments from track trajectory. Implemented neural network was able to reliably reconstruct missing segment positions with accuracy of $\approx 0.35$ wires, and lead to recovery of missing tracks with accuracy of $>99.8\%$.
