Visual Object Tracking Algorithm Based on Biological Visual Information Features and Few-Shot Learning

Zhang, Dawei; Yang, Tingting

doi:10.1155/2022/3422859

Cited by 6 publications

(3 citation statements)

References 26 publications

(28 reference statements)

Supporting

Mentioning

Contrasting

Order By: Relevance

“…Few-Shot Classification. Currently, many works have been proposed to address FSC [11][12][13][14][15][16][17][18][19], which can be mainly divided into three categories: initialization-based methods, metric-based methods, and hallucination-based methods.…”

Section: Related Workmentioning

confidence: 99%

Hybrid Fine-Tuning Strategy for Few-Shot Classification

Zhao

Zhang

et al. 2022

Computational Intelligence and Neuroscience

View full text Add to dashboard Cite

Few-shot classification aims to enable the network to acquire the ability of feature extraction and label prediction for the target categories given a few numbers of labeled samples. Current few-shot classification methods focus on the pretraining stage while fine-tuning by experience or not at all. No fine-tuning or insufficient fine-tuning may get low accuracy for the given tasks, while excessive fine-tuning will lead to poor generalization for unseen samples. To solve the above problems, this study proposes a hybrid fine-tuning strategy (HFT), including a few-shot linear discriminant analysis module (FSLDA) and an adaptive fine-tuning module (AFT). FSLDA constructs the optimal linear classification function under the few-shot conditions to initialize the last fully connected layer parameters, which fully excavates the professional knowledge of the given tasks and guarantees the lower bound of the model accuracy. AFT adopts an adaptive fine-tuning termination rule to obtain the optimal training epochs to prevent the model from overfitting. AFT is also built on FSLDA and outputs the final optimum hybrid fine-tuning strategy for a given sample size and layer frozen policy. We conducted extensive experiments on mini-ImageNet and tiered-ImageNet to prove the effectiveness of our proposed method. It achieves consistent performance improvements compared to existing fine-tuning methods under different sample sizes, layer frozen policies, and few-shot classification frameworks.

show abstract

Section: Related Workmentioning

confidence: 99%

Hybrid Fine-Tuning Strategy for Few-Shot Classification

Zhao

Zhang

et al. 2022

Computational Intelligence and Neuroscience

View full text Add to dashboard Cite

show abstract

“…There are several realworld uses for object tracking, including security, surveillance, autonomous driving, automated traffic management, biological image analysis, and intelligent robot control [10], [11]. The goal of object tracking, like that of the majority of computer vision systems, is to identify and extract the target item from a stream of images that the camera continuously records [12]- [14]. Better object tracking is made possible by faster image processing computation.…”

Section: Introductionmentioning

confidence: 99%

Field programmable gate array based moving object tracking system for robot navigation

Tran

Pham

et al. 2023

Bulletin EEI

View full text Add to dashboard Cite

This paper proposes a method in which an object tracking robot system is implemented on field programmable gate arrays (FPGAs). The OV7670 camera provides real-time object pictures to the system. To improve picture quality, images are put via the median filter phase. The item is distinguished from the backdrop based on color (red), after which it is subjected to a mathematical morphological approach of filtering to eliminate noise. To send the robot control signals, the object's (new) coordinates are found. In this method, the median filter, color separation, hardware IP cores, and morphological filter are all part of the embedded system on FPGA. Through the direct memory access (DMA) controller, these cores may communicate and perform high-speed pipeline computing at higher data rates. The entire system is executed in real-time on Xilinx's spartan-6 FPGA KIT. The results show practical and promise.

show abstract

“…With the continuous updating and development of technology, specific moving target tracking can be applied in many scenarios, such as industrial automation identification 1,2 , video surveillance 3,4 , and robot vision 5 . In the current related research, single-target tracking is one of the research hotspots for specific moving target tracking 6,7 , which aims to accurately and continuously localize the specific single object.…”

Section: Introductionmentioning

confidence: 99%

Tracking moving targets of interest-based deep learning

Shangguan,

Gong,

Chen

et al. 2023

5th International Conference on Information Science, Electrical, and Automation Engineering (ISEAE 2023)

View full text Add to dashboard Cite

Tracking specific target is an important research area in machine vision. When tracking specific target encounters the same local features with light and dark variations, convolutional neural networks based on deep learning require large data sets to extract target feature information. To address these problems, this paper proposes a target of interest tracking scheme based on specific object imaging. The method allows for fast capture and processing of important information about the target of interest. In addition, small data sets are used to track the target and extract the target of interest efficiently. The target of interest can also be accurately tracked when the background and light changes. The results show that our method reduces the processing of redundant information, reduces the storage overhead, and tracks the target accurately.

show abstract

Visual Object Tracking Algorithm Based on Biological Visual Information Features and Few-Shot Learning

Cited by 6 publications

References 26 publications

Hybrid Fine-Tuning Strategy for Few-Shot Classification

Hybrid Fine-Tuning Strategy for Few-Shot Classification

Field programmable gate array based moving object tracking system for robot navigation

Tracking moving targets of interest-based deep learning

Contact Info

Product

Resources

About