Robust face recognition for occluded real‐world images using constrained probabilistic sparse network

Ma, Xiang; Ma, Qinqin; Ma, Qian; Han, Xiao

doi:10.1049/ipr2.12414

Cited by 5 publications

(4 citation statements)

References 61 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…For strengthening the anti-interference performance of FR models against occlusions, Ma et al presented a FR method on the ground of second-order degree constraints and verified its effectiveness. It was found that compared with non-deep learning methods, this method possesses the best recognition rate [15].…”

Section: Related Workmentioning

confidence: 99%

The Application of Improved Scale Invariant Feature Transformation Algorithm in Facial Recognition

Cong

2024

IJACSA

View full text Add to dashboard Cite

Currently, face recognition models suffer from insufficient accuracy, stability, and computational efficiency. To address this issue, an improved feature extraction algorithm on the ground of Haar wavelet features and scale invariant feature transformation algorithm is proposed. In addition, the study also combines this algorithm with deep belief networks to construct an improved facial recognition model. The effectiveness of the proposed improved feature extraction algorithm was verified, and it was found that the recognition accuracy of the algorithm was 94.2%, which is better than other comparative algorithms. In addition, the study also conducted empirical analysis on the improved facial recognition model and found that the recognition accuracy of the model was 0.92, and the feature matching time was 2.6 seconds, which was better than other comparative models in terms of performance. On the ground of the above results, the proposed facial recognition model has significantly improved recognition accuracy and efficiency compared to traditional models. It can provide theoretical reference for improving the universality of facial recognition applications in different fields.

show abstract

Section: Related Workmentioning

confidence: 99%

The Application of Improved Scale Invariant Feature Transformation Algorithm in Facial Recognition

Cong

2024

IJACSA

View full text Add to dashboard Cite

show abstract

“…Wang et al [ 16 ] proposed an object shape feature extraction approach called slope difference distribution (SDD), which extracts features of shape as a sparse representation and utilizes the detected SDD features of all shape models and the minimum distance between SDD features for object recognition. Ma et al [ 17 ] proposed a robust face recognition approach based on a sparse network with limited probability, built a sparse image network with limited probability, and acquired the overall training images from a global perspective for recognition. Heo et al [ 18 ] proposed an occlusion-aware spatial attention transformer (OSAT) architecture based on a visual transformer (ViT), CutMix strengthening, and occlusion mask predictor (OMP) to solve the occlusion problem.…”

Section: Related Workmentioning

confidence: 99%

Recognition of Occluded Goods under Prior Inference Based on Generative Adversarial Network

Cao

Xie

Liu

et al. 2023

Sensors

View full text Add to dashboard Cite

Aiming at the recognition of intelligent retail dynamic visual container goods, two problems that lead to low recognition accuracy must be addressed; one is the lack of goods features caused by the occlusion of the hand, and the other is the high similarity of goods. Therefore, this study proposes an approach for occluding goods recognition based on a generative adversarial network combined with prior inference to address the two abovementioned problems. With DarkNet53 as the backbone network, semantic segmentation is used to locate the occluded part in the feature extraction network, and simultaneously, the YOLOX decoupling head is used to obtain the detection frame. Subsequently, a generative adversarial network under prior inference is used to restore and expand the features of the occluded parts, and a multi-scale spatial attention and effective channel attention weighted attention mechanism module is proposed to select fine-grained features of goods. Finally, a metric learning method based on von Mises–Fisher distribution is proposed to increase the class spacing of features to achieve the effect of feature distinction, whilst the distinguished features are utilized to recognize goods at a fine-grained level. The experimental data used in this study were all obtained from the self-made smart retail container dataset, which contains a total of 12 types of goods used for recognition and includes four couples of similar goods. Experimental results reveal that the peak signal-to-noise ratio and structural similarity under improved prior inference are 0.7743 and 0.0183 higher than those of the other models, respectively. Compared with other optimal models, mAP improves the recognition accuracy by 1.2% and the recognition accuracy by 2.82%. This study solves two problems: one is the occlusion caused by hands, and the other is the high similarity of goods, thus meeting the requirements of commodity recognition accuracy in the field of intelligent retail and exhibiting good application prospects.

show abstract

“…ProCRC effectively uses training samples from all classes to deduce 2 the class label of the test sample. Ma et al [6] proposed a constrained probabilistic sparse representation network, which first introduced the second-order gradient constraint in the probabilistic sparse representation network. Zheng et al [7] proposed a feature extraction method based on structural elements, which can capture local and contextual information and adaptively fuse multiple features to improve recognition accuracy.…”

Section: Introductionmentioning

confidence: 99%

Occluded Face Recognition with a Novel Model Incorporating Block Diagonal Structure and Feature Consistency

Guo

Huo

2023

Preprint

Self Cite

View full text Add to dashboard Cite

The nuclear norm matrix regression method is effective for continuous occlusion in face recognition. However, the existing method only considers low-rank structural information and ignores correlation between sample image representations. To effectively solve these issues, we propose a novel occluded face recognition model. The model enhances differences between categories using a strict 0–1 block diagonal structure. It also improves feature representation consistency within the same category with a local preservation term. The introduction of these two terms enables the model to obtain more discriminative representation coefficients. The experimental results on the Extended Yale B, AR, and LFW databases demonstrate that the proposed method has better recognition performance for occluded face recognition than comparative methods.

show abstract

Robust face recognition for occluded real‐world images using constrained probabilistic sparse network

Cited by 5 publications

References 61 publications

The Application of Improved Scale Invariant Feature Transformation Algorithm in Facial Recognition

The Application of Improved Scale Invariant Feature Transformation Algorithm in Facial Recognition

Recognition of Occluded Goods under Prior Inference Based on Generative Adversarial Network

Occluded Face Recognition with a Novel Model Incorporating Block Diagonal Structure and Feature Consistency

Contact Info

Product

Resources

About