Singular Value Fine-tuning: Few-shot Segmentation requires Few-parameters Fine-tuning

Sun, Yu; Chen, Qiang; He, Xiaofeng; Wang, Jian; Feng, Hailin; Han, Junyu; Ding, Errui; Cheng, Jun; Li, Zechao; Wang, Jingdong

doi:10.48550/arxiv.2206.06122

Cited by 4 publications

(4 citation statements)

References 33 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…With a Resnet-50 backbone, we achieve results comparable to other methods with similar number of trainable parameters. The most recent work of Sun et al 48 , which achieves state-of-the-art with a Resnet-50, significantly increases the memory requirements because, according to the authors, it requires 128G for a batch 8 (16G for one image), which is significantly higher than any other few-shot semantic segmentation technique.…”

Section: Discussionmentioning

confidence: 99%

“…See appendix for full-sized table. SVF48 46.87 53.80 48.43 44.78 48.47 52.25 57.83 51.97 53.41 53.87 BAM 12 43.41 50.59 47.49 43.42 46.23 49.26 54.20 51.63 49.55 51.16 Baseline-HSNet 36.3 43.1 38.7 39.2 39.2 43.3 51.3 48.2 45.0 46.9 Ours 42.15 53.22 49.05 48.08 48.12 47.50 59.14 53.19 51.16 52.75 HSNet 37.2 44.1 42.4 41.3 41.2 45.9 53.0 51.8 47.1 49.5 Ours 45.48 56.47 51.74 49.84 50.88 48.87 61.10 55.58 54.03 54.90…”

mentioning

confidence: 99%

See 1 more Smart Citation

Transductive Meta-Learning with Enhanced Feature Ensemble for Few-shot Semantic Segmentation

Karimi,

POULLIS

2023

Preprint

View full text Add to dashboard Cite

This paper addresses few-shot semantic segmentation and proposes a novel transductive end-to-end method that overcomes three key problems affecting performance. First, we present a novel ensemble of visual features learned from pretrained classification and semantic segmentation networks with the same architecture. Our approach leverages the varying dis-criminative power of these networks, resulting in visual features that capture rich and diverse information at different depths. Secondly, the pretrained semantic segmentation network serves as a base class extractor, which significantly reduces false positives. Thirdly, a transductive meta-learning method addresses the prediction by first learning the relationship between labeled and unlabeled data points with matching support foreground to query features (intra-class similarity) and then applying this knowledge to predict on the unlabeled query image (intra-object similarity), which simultaneously learns propagation and false positive suppression. To evaluate our method, we performed experiments on benchmark datasets, and the results demonstrate significant improvement, with minimal trainable parameters 2.98M. Specifically, using Resnet-101, we achieve state-of-the-art performance for both 1-shot and 5-shot Pascal-5i , as well as on 1-shot and 5-shot COCO-20i .

show abstract

Section: Discussionmentioning

confidence: 99%

mentioning

confidence: 99%

Transductive Meta-Learning with Enhanced Feature Ensemble for Few-shot Semantic Segmentation

Karimi,

POULLIS

2023

Preprint

View full text Add to dashboard Cite

show abstract

“…SSA outperformed baseline state-of-the-art methods for diverse object localization tasks. Sun et al [16] introduced a novel approach to few-shot segmentation (FSS) named Singular Value Fine-tuning (SVF), which addressed the overfitting issue in FSS. In FSS, the task is to segment novel class objects with only a few densely annotated samples.…”

Section: A Segmentationmentioning

confidence: 99%

Neural Inference Search for Multiloss Segmentation Models

Slade

Zhang

Huang

et al. 2024

IEEE Trans. Neural Netw. Learning Syst.

View full text Add to dashboard Cite

Semantic segmentation is vital for many emerging surveillance applications, but current models cannot be relied upon to meet the required tolerance, particularly in complex tasks that involve multiple classes and varied environments. To improve performance, we propose a novel algorithm, Neural Inference Search (NIS), for hyperparameter optimisation pertaining to established deep learning segmentation models in conjunction with a new multi-loss function. It incorporates three novel search behaviours, i.e. Maximised Standard Deviation Velocity Prediction, Local Best Velocity Prediction, and n-dimensional Whirlpool Search. The first two behaviours are exploratory, leveraging Long Short-Term Memory (LSTM)-(Convolutional Neural Network) CNN based velocity predictions, while the third employs n-dimensional matrix rotation for local exploitation.A scheduling mechanism is also introduced in NIS to manage the contributions of these three novel search behaviours in stages. NIS optimises learning and multi-loss parameters simultaneously. Compared with state-of-the-art segmentation methods and those optimised with other well-known search algorithms, NISoptimised models show significant improvements across multiple performance metrics on five segmentation datasets. NIS also reliably yields better solutions as compared with a variety of search methods for solving numerical benchmark functions.

show abstract

“…Y. Sun et al [31] proposed a model fine-tuning method based on singular value decomposition, which can perform fast model fine-tuning by using a very small number of parameters, thus achieving the segmentation task with few samples. Z. Li et al [32]proposed a knowledge-guided approach for few-sample image recognition.…”

Section: Related Workmentioning

confidence: 99%

Global enhancement network underwater archaeology scene parsing method

Pan,

Jia,

Cai

2023

Robotica

View full text Add to dashboard Cite

Underwater archaeology is of great significance for historical and cultural transmission and preservation of underwater heritage, but it is also a challenging task. Underwater heritage is located in an environment with high sediment content, objects are mostly buried, and the water is turbid, resulting in some of the features of objects missing or blurred, making it difficult to accurately identify and understand the semantics of various objects in the scene. To tackle these issues, this paper proposes a global enhancement network (GENet) underwater scene parsing method. We introduce adaptive dilated convolution by adding an extra regression layer, which can automatically deduce adaptive dilated coefficients according to the different scene objects. In addition, considering the easy confusion in the process of fuzzy feature classification, an enhancement classification network is proposed to increase the difference between various types of probabilities by reducing the loss function. We verified the validity of the proposed model by conducting numerous experiments on the Underwater Shipwreck Scenes (USS) dataset. We achieve state-of-the-art performance compared to the current state-of-the-art algorithm under three different conditions: conventional, relic semi-buried, and turbidified water quality. The experimental results show that the proposed algorithm performs best in different situations. To verify the generalizability of the proposed algorithm, we conducted comparative experiments on the current publicly available Cityscapes, ADE20K, and the underwater dataset SUIM. The experimental results show that this paper achieves good performance on the public dataset, indicating that the proposed algorithm is generalizable.

show abstract

Singular Value Fine-tuning: Few-shot Segmentation requires Few-parameters Fine-tuning

Cited by 4 publications

References 33 publications

Transductive Meta-Learning with Enhanced Feature Ensemble for Few-shot Semantic Segmentation

Transductive Meta-Learning with Enhanced Feature Ensemble for Few-shot Semantic Segmentation

Neural Inference Search for Multiloss Segmentation Models

Global enhancement network underwater archaeology scene parsing method

Contact Info

Product

Resources

About