Face photo-sketch recognition using local and global texture descriptors

Galea, Christian; Farrugia, Reuben A.

doi:10.1109/eusipco.2016.7760647

Cited by 23 publications

(29 citation statements)

References 19 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…2) Comparison to the state-of-the-art methods For the UoM-SGFSv2 set A and B, we report the performance of our DAEN with the EP(+PCA) [47], ET(+PCA) [48], D-RS+CBR [49], LGMS [50], DEEPS [9], SP-Net [36], Identity-aware CycleGAN [29], and Transfer deep feature learning [37] methods in Table 5, LGMS are traditional inter-modality methods, the Identityaware CycleGAN is a deep-learning intra-modality method, the DEEPS, SP-Net, and Transfer deep feature learning methods are deep-learning inter-modality methods. As we can see from Table 5, Table 6 and Figure 10, Figure 11, the proposed DAEN achieves the highest accuracy, and the performance of all methods is lower on the more challenging UoM-SGFSv2 set A.…”

Section: Results and Analysis 1) Ablation Studymentioning

confidence: 99%

Domain Alignment Embedding Network for Sketch Face Recognition

et al. 2021

View full text Add to dashboard Cite

Section: Results and Analysis 1) Ablation Studymentioning

confidence: 99%

Domain Alignment Embedding Network for Sketch Face Recognition

et al. 2021

View full text Add to dashboard Cite

“…In this database, they built a model to reverse the process. Galea et al [4] proposed the method that extracts multi scale local binary pattern descriptors from the overlapping patches for sketch image. Global and local texture information can be used for face photo sketch recognition.…”

Section: Related Workmentioning

confidence: 99%

Face Photo Recognition using Sketch Image for Security System

2019

IJITEE

View full text Add to dashboard Cite

Because of expanding requests in application regions, for example, validation of law requirement, video reconnaissance, banking and access to security frameworks, programmed face acknowledgment has pulled in extraordinary consideration as of late. The acknowledgment of the face has ended up being one of the present essential applications. It is utilized fundamentally for computerized amusement, security and authorization. Face acknowledgment can be for the entire or fractional face, the entire face for the entire face identification and explicit highlights for the incomplete face discovery. The examination in most criminal cases relies upon the portrayals attracted by the observers ' depiction. The programmed recovery of photographs from the police mug-shot database that coordinate this depiction can conceivably diminish the quantity of suspects. Fundamentally, this procedure relies upon the onlooker and precision of the sketch craftsman's memory of the face to catch those subtleties. It can help scientists viably find or diminish potential suspects. By and large, be that as it may, a speculate's photograph picture isn't accessible and the best substitute is frequently a sketch drawing dependent on an observer's memory. We present another photograph recuperation framework utilizing face draws by changing a photograph picture into a sketch, fundamentally lessening the distinction among photograph and sketch, enabling the two to coordinate successfully.

show abstract

“…using a linear combination of images, the Eigen-patches (EP) extension [15] performing synthesis at a local level, and the Bayesian framework in [16] that considers relationships among neighbouring patches for model construction. A more thorough review of FH algorithms may be found in [4], [13], [15]- [17]. State-of-the-art inter-modality methods that learn or extract modality-invariant features include the D-RS approach [2], [18] that compares SIFT and MLBP descriptors extracted from images that are convolved with three filters, the CBR method [19], which compares MLBP features extracted from individual facial components, the FaceSketchID system in [12] which fuses D-RS with CBR, and the recent LGMS method [4] that compares MLBP features extracted from log-Gabor-filtered images using the spearman rank-order correlation coefficient.…”

Section: Related Workmentioning

confidence: 99%

“…A more thorough review of FH algorithms may be found in [4], [13], [15]- [17]. State-of-the-art inter-modality methods that learn or extract modality-invariant features include the D-RS approach [2], [18] that compares SIFT and MLBP descriptors extracted from images that are convolved with three filters, the CBR method [19], which compares MLBP features extracted from individual facial components, the FaceSketchID system in [12] which fuses D-RS with CBR, and the recent LGMS method [4] that compares MLBP features extracted from log-Gabor-filtered images using the spearman rank-order correlation coefficient. Other methods and further information can be found in [4], [6], [12], [13], [15], [19]- [23].…”

Section: Related Workmentioning

confidence: 99%

“…Photos and sketches populate the gallery and probe sets, respectively, and the gallery is extended with the photos of 1521 subjects to simulate the extensive mug-shot galleries maintained by law-enforcement agencies, obtained from the MEDS-II, 4 FRGC v2.0, 5 Multi-PIE [44], and FEI 6 databases. 4 Available at: http://www.nist.gov/itl/iad/ig/sd32.cfm 5 Available at: http://www.nist.gov/itl/iad/ig/frgc.cfm 6 Available at: http://fei.edu.br/∼cet/facedatabase.html As shown in Table I, the FRSs are generally inferior to the intra-modality methods, which in turn typically lag behind the performance of the inter-modality approaches. The only exception is CBR, whose poor performance is likely a result of being designed to operate on software-generated sketches.…”

Section: System Fusionmentioning

confidence: 99%

See 1 more Smart Citation

Forensic Face Photo-Sketch Recognition Using a Deep Learning-Based Architecture

Galea

Farrugia

2017

IEEE Signal Process. Lett.

Self Cite

View full text Add to dashboard Cite

Numerous methods that automatically identify subjects depicted in sketches as described by eyewitnesses have been implemented, but their performance often degrades when using real-world forensic sketches and extended galleries that mimic law enforcement mug-shot galleries. Moreover, little work has been done to apply deep learning for face photo-sketch recognition despite its success in numerous application domains including traditional face recognition. This is primarily due to the limited number of sketch images available, which are insufficient to robustly train large networks. This letter aims to tackle these issues with the following contributions: 1) a state-of-the-art model pre-trained for face photo recognition is tuned for face photo-sketch recognition by applying transfer learning, 2) a three-dimensional morphable model is used to synthesise new images and artificially expand the training data, allowing the network to prevent over-fitting and learn better features, 3) multiple synthetic sketches are also used in the testing stage to improve performance, and 4) fusion of the proposed method with a state-of-the-art algorithm is shown to further boost performance. An extensive evaluation of several popular and state-of-the-art algorithms is also performed using publicly available datasets, thereby serving as a benchmark for future algorithms. Compared to a leading method, the proposed framework is shown to reduce the error rate by 80.7% for viewed sketches and lowers the mean retrieval rank by 32.5% for real-world forensic sketches.

show abstract

Face photo-sketch recognition using local and global texture descriptors

Cited by 23 publications

References 19 publications

Domain Alignment Embedding Network for Sketch Face Recognition

Domain Alignment Embedding Network for Sketch Face Recognition

Face Photo Recognition using Sketch Image for Security System

Forensic Face Photo-Sketch Recognition Using a Deep Learning-Based Architecture

Contact Info

Product

Resources

About