Local binary pattern-based discriminant graph construction for dimensionality reduction with application to face recognition

Yang, Bo; Li, Qian‐Zhong

doi:10.1007/s11042-019-7518-3

Cited by 2 publications

(2 citation statements)

References 30 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…Currently, PCA and LDA remain the two most popular DR approaches. Yang and Li proposed a novel algorithm using LBP-based discriminant graph construction for reducing dimensionality, which is applicable to face recognition [11].…”

Section: Literature Reviewmentioning

confidence: 99%

“…Figs. [8][9][10][11] indicate that although they were converted to lower resolution, the test data acquired using a high-resolution camera still realized good accuracy with an average RR of four face objects higher than 50%. A comparison of the data also displays similarities in the local and real-time data files, whereby the smallest recognition time interval of 0.1 seconds produced the lowest accuracy value.…”

mentioning

confidence: 99%

See 1 more Smart Citation

Multi-Object Face Recognition Using Local Binary Pattern Histogram and Haar Cascade Classifier on Low-Resolution Images

Isnanto

Rochim

Eridani

et al. 2021

Int. j. eng. technol. innov.

View full text Add to dashboard Cite

This study aims to build a face recognition prototype that can recognize multiple face objects within one frame. The proposed method uses a local binary pattern histogram and Haar cascade classifier on low-resolution images. The lowest data resolution used in this study was 76 × 76 pixels and the highest was 156 × 156 pixels. The face images were preprocessed using the histogram equalization and median filtering. The face recognition prototype proposed successfully recognized four face objects in one frame. The results obtained were comparable for local and real-time stream video data for testing. The RR obtained with the local data test was 99.67%, which indicates better performance in recognizing 75 frames for each object, compared to the 92.67% RR for the real-time data stream. In comparison to the results obtained in previous works, it can be concluded that the proposed method yields the highest RR of 99.67%.

show abstract

Section: Literature Reviewmentioning

confidence: 99%

mentioning

confidence: 99%