Aiming at solving the problem of missed detection and low accuracy in detecting traffic signs in the wild, an improved method of YOLOv8 is proposed. Firstly, combined with the characteristics of small target objects in the actual scene, this paper further adds blur and noise operation. Then, the asymptotic feature pyramid network (AFPN) is introduced to highlight the influence of key layer features after feature fusion, and simultaneously solve the direct interaction of non-adjacent layers. Experimental results on the TT100K dataset show that compared with the YOLOv8, the detection accuracy and recall are higher.
No abstract
Part-based adaptive appearance model has been extensively used in increasingly popular discriminative trackers. The main problem of these methods is the stability plasticity dilemma. Embedding holistic appearance information in the part-based appearance model which is learned online to alleviate this problem is proposed. Specifically, the object is represented by sparse multi-scale Haar-like features and the appearance model is constructed with a naive Bayes classifier. Unlike the conventional methods, the classifier is trained by positive and negative samples that are weighted according to their similarity with the holistic appearance model, which is kept constant during the updating procedure. The constant holistic appearance information providing some constraints when updating the part-based appearance model makes the tracker more stable. The online updating procedure of the part-based appearance model makes the tracker adaptive enough to appearance changes. Experimental results demonstrate the superior performance of the proposed method compared with several state-ofart algorithms.Introduction: Recently, algorithms which regard tracking as a binary classification task and build the appearance model via online updated classifiers have greatly increased in popularity. These methods [1][2][3][4] representing the target as a set of Haar-like features [5] have achieved promising results in handling appearance changes caused by pose changes, illumination changes, occlusion and irregular object motion. However, they suffer the stability plasticity dilemma [6]. Grabner et al.[1] propose an online boosting algorithm to select features for tracking. This self-learning scheme could lead to a drifting problem by noisy or misclassified samples. Grabner et al. [2] propose an online semi-supervised boosting method to alleviate the drifting problem by only labelling the samples in the first frame and labelling samples of subsequent frames in a soft labelling scheme. This method suffers less drifting, but has limited adaptation capabilities. Babenko et al.[3] introduce multiple instance learning into online tracking, where samples are considered within positive and negative bags to alleviate the noise of hard labelling. This sample labelling mechanism relieves the drifting problem to a certain extent, but still cannot avoid it due to the online updating without extra appearance constraints.In this Letter, we introduce holistic appearance information into the online updating procedure of the part-based appearance model. In particular, the object is represented by sparse multi-scale Haar-like features and the appearance model is constructed with a naive Bayes classifier. Unlike the conventional methods [1-4], the training positive and negative samples are weighted according to their similarity with a holistic appearance model that is maintained at constant during the whole tracking process. This approach relieves the drifting problem significantly and keeps the tracker adaptive enough to appearance changes. Experimental results...
scite is a Brooklyn-based organization that helps researchers better discover and understand research articles through Smart Citations–citations that display the context of the citation and describe whether the article provides supporting or contrasting evidence. scite is used by students and researchers from around the world and is funded in part by the National Science Foundation and the National Institute on Drug Abuse of the National Institutes of Health.
customersupport@researchsolutions.com
10624 S. Eastern Ave., Ste. A-614
Henderson, NV 89052, USA
Copyright © 2024 scite LLC. All rights reserved.
Made with 💙 for researchers
Part of the Research Solutions Family.