paper-with-me

홈 › Papers

Eliminate Deviation with Deviation for Data Augmentation and a General Multi-modal Data Learning Method

2021-01-21 · Yunpeng Gong, Liqing Huang, Lifei Chen

One of the challenges of computer vision is that it needs to adapt to color deviations in changeable environments. Therefore, minimizing the adverse effects of color deviation on the prediction is one of the main goals of vision task. Current solutions focus on using generative models to augment training data to enhance the invariance of input variation. However, such methods often introduce new noise, which limits the gain from generated data. To this end, this paper proposes a strategy eliminate deviation with deviation, which is named Random Color Dropout (RCD). Our hypothesis is that if there are color deviation between the query image and the gallery image, the retrieval results of some examples will be better after ignoring the color information. Specifically, this strategy balances the weights between color features and color-independent features in the neural network by dropouting partial color information in the training data, so as to overcome the effect of color devitaion. The proposed RCD can be combined with various existing ReID models without changing the learning strategy, and can be applied to other computer vision fields, such as object detection. Experiments on several ReID baselines and three common large-scale datasets such as Market1501, DukeMTMC, and MSMT17 have verified the effectiveness of this method. Experiments on Cross-domain tests have shown that this strategy is significant eliminating the domain gap. Furthermore, in order to understand the working mechanism of RCD, we analyzed the effectiveness of this strategy from the perspective of classification, which reveals that it may be better to utilize many instead of all of color information in visual tasks with strong domain variations.

📄 PDF Abstract BibTeX arXiv:2101.08533

Code (1)

finger-monkey/Data-Augmentation 공식 구현 pytorch

Tasks

Adversarial DefenseData Augmentationobject-detectionObject DetectionPerson Re-IdentificationRetrieval

Methods 이 논문이 사용한 방법론

Random Grayscale Random Grayscale is an image data augmentation that converts an image to grayscale with probability $p$.

Similar Papers 제목 키워드 기반

Graph Out-of-Distribution Generalization with Controllable Data Augmentation

2023-08-16 · Bin Lu, Xiaoying Gan, Ze Zhao, Shiyu Liang 외

Graph Neural Network (GNN) has demonstrated extraordinary performance in classifying graph properties. However, due to the selection bias of training and testing data (e.g., training on small graphs and testing on large …

Data AugmentationGraph ClassificationGraph Neural NetworkOut-of-Distribution Generalization+1

Window-Based Distribution Shift Detection for Deep Neural Networks

2022-10-19 · NeurIPS 2023 11

To deploy and operate deep neural models in production, the quality of their predictions, which might be contaminated benignly or manipulated maliciously by input distributional deviations, must be monitored and assessed…

DEeR: Deviation Eliminating and Noise Regulating for Privacy-preserving Federated Low-rank Adaptation

2024-10-16 · Meilu Zhu, Axiu Mao, Jun Liu, Yixuan Yuan

Integrating low-rank adaptation (LoRA) with federated learning (FL) has received widespread attention recently, aiming to adapt pretrained foundation models (FMs) to downstream medical tasks via privacy-preserving decent…

Federated LearningPrivacy Preserving

Prompt-Free Conditional Diffusion for Multi-object Image Augmentation

2025-07-08 · Haoyu Wang, Lei Zhang, Wei Wei, Chen Ding 외

Diffusion models has underpinned much recent advances of dataset augmentation in various computer vision tasks. However, when involving generating multi-object images as real scenarios, most existing methods either rely …

DiversityDomain GeneralizationImage Augmentation

Beyond Normal References: Discriminative Few-Shot Anomaly Detection

2026-05-22 · Huan Wang, Jun Shen, Jun Yan, Guansong Pang arxiv

This paper considers a practical few-shot anomaly detection (FSAD) setting, termed discriminative FSAD, where a limited number of both normal and anomalous examples are available as references during inference. Existing …

Anomaly Detection