Temporal Speciation Network for Few-Shot Object Detection

Zhao, Xiaowei; Liu, Xianglong; Ma, Yuqing; Bai, Shihao; Shen, Yifan; Hao, Zeyu; Liu, Aishan

doi:10.1109/tmm.2023.3234368

Cited by 9 publications

(3 citation statements)

References 55 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…In Table 2, we compare our method with two‐branch detectors including Meta R‐CNN [29], Attention‐RPN [46], FsDetView [39], Dense Relation Distillation with Context‐aware Aggregation Network [31], CME [14], Transformation Invariant Principle [47], Meta‐DETR [36], Few‐Shot Object Detection with Universal Prototypes [11], Query Adaptive Few‐Shot Object Detection [48], Generate Detectors [49], Meta Faster R‐CNN [50], Intra‐Support Attention Module and the Query‐Support Attention Module [51], CAReD [32] and Kernelized Few‐Shot Object Detection [52], which are mata‐learning‐based methods and single‐branch detectors including TFA [33], MPSR [30], Semantic Relation Reasoning for Shot‐Stable Few‐Shot Object Detection [53], FSCE [13], Cooperating Region Proposal Network’s (CoRPNs) + Hallucination [54], Singular Value Decomposition [55], Few‐Shot Object Detection via Association and Discrimination [56], Decoupled Faster Region based Convolutional Neural Network [57] and TeSNet [35], which are fine‐tuning‐based methods. And it can be seen that our method has a great improvement over other state‐of‐the‐art methods.…”

Section: Resultsmentioning

confidence: 99%

“…When the number of support sets is large, it needs to be fine‐tuned many times, so the scheme is cumbersome. Temporal Speciation Network (TesNet) [35], imitating the natural evolution which relies on inheritation and mutation, improves the diversity and rationality of positive proposal generation. Existing fine‐tuning‐based methods believe that the ability of the network to distinguish between positive and negative samples is consistent in the novel and base classes.…”

Section: Related Workmentioning

confidence: 99%

See 1 more Smart Citation

Few‐shot object detection via class encoding and multi‐target decoding

Guo

Yang

Wei

et al. 2023

IET Cyber-Syst and Robotics

View full text Add to dashboard Cite

The task of few-shot object detection is to classify and locate objects through a few annotated samples. Although many studies have tried to solve this problem, the results are still not satisfactory. Recent studies have found that the class margin significantly impacts the classification and representation of the targets to be detected. Most methods use the loss function to balance the class margin, but the results show that the loss-based methods only have a tiny improvement on the few-shot object detection problem. In this study, the authors propose a class encoding method based on the transformer to balance the class margin, which can make the model pay more attention to the essential information of the features, thus increasing the recognition ability of the sample. Besides, the authors propose a multi-target decoding method to aggregate RoI vectors generated from multi-target images with multiple support vectors, which can significantly improve the detection ability of the detector for multi-target images. Experiments on Pascal visual object classes (VOC) and Microsoft Common Objects in Context datasets show that our proposed Few-Shot Object Detection via Class Encoding and Multi-Target Decoding significantly improves upon baseline detectors (average accuracy improvement is up to 10.8% on VOC and 2.1% on COCO), achieving competitive performance. In general, we propose a new way to regulate the class margin between support set vectors and a way of feature aggregation for images containing multiple objects and achieve remarkable results. Our method is implemented on mmfewshot, and the code will be available later.

show abstract

Section: Resultsmentioning

confidence: 99%

Section: Related Workmentioning

confidence: 99%

Few‐shot object detection via class encoding and multi‐target decoding

Guo

Yang

Wei

et al. 2023

IET Cyber-Syst and Robotics

View full text Add to dashboard Cite

show abstract

“…When these two novel elements were combined, they demonstrated that logit mimicking could perform better than feature imitation for the first time, and that the main cause of logit mimicking years-long underperformance is the lack of localization distillation [17]. Researchers introduced Temporal Speciation Network (TeSNet), a straightforward but efficient few-shot object detection framework with changing training that increased the diversity and rationality of positive suggestion creation [18].…”

Section: Literature Reviewmentioning

confidence: 99%

YOLO/SBHK-Net: Occlusion Aware Robot Vision Neural Network with Compound Model Scaling and Re-parameterized Convolution

Banerjee,

Kabir

2024

Preprint

View full text Add to dashboard Cite

Robot Vision is the technique of enabling robots to process visual data from the environment by utilizing a combination of camera hardware and computer algorithms. Advanced deep neural networks have significantly played a vital role in indulging robots to make more sense out of complex visual data at different circumstances, especially in object detection and continuous tracking. In this research, we initiated a unique and cutting-edge backbone neural network for the conventional YOLO algorithm which we named as SBHK-Net. The network boosted up the performance of the existing YOLO algorithm drastically which manifests a strong potential of improving tracking and recognition accuracies of other conventional algorithms in the robot vision industry as well. It has the greatest accuracy 59.2% AP among all known real-time object detectors with 30 FPS or above on GPU RTX3060, and it outperforms all other known object detectors in the range of 5 FPS to 160 FPS. We used YOLOv7 as our reference point for the core research. The transformer-based detector SWINL Cascade-Mask R-CNN (9.2 FPS A100, 53.9% AP) and the convolutional detector ConvNeXt-XL Cascade-Mask R-CNN (8.6 FPS A100, 55.2% AP) are both outperformed by the SBHK-Net core object detector (56 FPS RTX3060, 56.4% AP) in terms of speed and accuracy, respectively. In terms of speed and accuracy, it surpasses a number of other object detectors, including DINO-5scale-R50, ViT-Adapter-B, Scaled-YOLOv4, YOLOv5, DETR, Deformable DETR, and YOLOR and YoLOX. The source code of this research is available at https://github.com/ac005sheekar/SBHKNet.

show abstract

Class‐Relation Reasoning with Knowledge‐Transfer for Few‐Shot Object Detection

Feng,

Zhang,

Wang

et al. 2024

IEEJ Transactions Elec Engng

View full text Add to dashboard Cite

Few‐Shot Object Detection (FSOD) task involves accurately identifying target object classes using only a small set of labeled samples. Most of the current FSOD tasks independently predict class prototype features without considering class relationships and only rely on visual information. To address these challenges, we propose a novel Class‐relational Reasoning Method with Knowledge‐transfer (CRK‐Net), built on the meta‐learning‐based framework. Although data may be scarce, the semantic relationship between classes is invariant, Joint‐feature Fusion Module (JFM) are hence proposed to transfers the semantic information of different categories in the natural language world to integrate with visual information and produce multi‐modality embeddings. Some base classes and novel classes have similar features, so this can be borrowed by modeling the relationship between classes feature. Building upon the observation, we propose a Class‐relational Reasoning Module (CRM) to establish the correlations between categories and enhance prototype representations for each category. After passing through the JFM and CRM modules, a high‐quality class prototype is finally produced for subsequent regression and classification. Extensive experiments on PASCAL VOC demonstrate the effectiveness of our proposed method and provide a new scheme for fusing semantic and visual information. © 2024 Institute of Electrical Engineers of Japan. Published by Wiley Periodicals LLC.

show abstract

Temporal Speciation Network for Few-Shot Object Detection

Cited by 9 publications

References 55 publications

Few‐shot object detection via class encoding and multi‐target decoding

Few‐shot object detection via class encoding and multi‐target decoding

YOLO/SBHK-Net: Occlusion Aware Robot Vision Neural Network with Compound Model Scaling and Re-parameterized Convolution

Class‐Relation Reasoning with Knowledge‐Transfer for Few‐Shot Object Detection

Contact Info

Product

Resources

About