Compared to a leading commercial face recognition system, LFDA offers substantial improvements in matching forensic sketches to the corresponding face images and leads to state-of-the-art accuracys when matching viewed sketches.
Abstract:
The problem of matching a forensic sketch to a gallery of mug shot images is addressed in this paper. Previous research in sketch matching only offered solutions to matching highly accurate sketches that were drawn while looking at the subject (viewed sketches). Forensic sketches differ from viewed sketches in that they are drawn by a police sketch artist using the description of the subject provided by an eyewitness. To identify forensic sketches, we present a framework called local feature-based discriminant analysis (LFDA). In LFDA, we individually represent both sketches and photos using SIFT feature descriptors and multiscale local binary patterns (MLBP). Multiple discriminant projections are then used on partitioned vectors of the feature-based representation for minimum distance matching. We apply this method to match a data set of 159 forensic sketches against a mug shot gallery containing 10,159 images. Compared to a leading commercial face recognition system, LFDA offers substantial improvements in matching forensic sketches to the corresponding face images. We were able to further improve the matching performance using race and gender information to reduce the target gallery size. Additional experiments demonstrate that the proposed framework leads to state-of-the-art accuracys when matching viewed sketches.
TL;DR: This work proposes a Multi-view Discriminant Analysis (MvDA) approach, which seeks for a single discriminant common space for multiple views in a non-pairwise manner by jointly learning multiple view-specific linear transforms.
TL;DR: This paper uses Partial Least Squares to linearly map images in different modalities to a common linear subspace in which they are highly correlated, and forms a generic intermediate subspace comparison framework for multi-modal recognition.
TL;DR: A generic HFR framework is proposed in which both probe and gallery images are represented in terms of nonlinear similarities to a collection of prototype face images, and Random sampling is introduced into the H FR framework to better handle challenges arising from the small sample size problem.
TL;DR: This paper proposes a method to learn a discriminant face descriptor (DFD) in a data-driven way and applies it to the heterogeneous (cross-modality) face recognition problem and learns DFD in a coupled way to reduce the gap between features of heterogeneous face images to improve the performance of this challenging problem.
TL;DR: A new face descriptor based on coupled information-theoretic encoding is used to capture discriminative local face structures and to effectively match photos and sketches by reducing the modality gap at the feature extraction stage.
TL;DR: This work proposes to generate a realistic face image from the composite sketch using a hybrid subspace method and then build an illumination tolerant correlation filter which can recognize the person under different illumination variations from a surveillance video footage.
TL;DR: A new database that includesRotating head videos of 259 subjects; 250 hand-drawn face sketches of 50 subjects; and this is the only currently available database that has a large number of face sketches drawn by multiple artists is described.
Q1. What contributions have the authors mentioned in the paper "Matching forensic sketches to mug shot photos" ?
The problem of matching a forensic sketch to a gallery of mug shot images is addressed in this paper. Forensic sketches differ from viewed sketches in that they are drawn by a police sketch artist using the description of the subject provided by an eyewitness. To identify forensic sketches, the authors present a framework called local feature-based discriminant analysis ( LFDA ). In LFDA, the authors individually represent both sketches and photos using SIFT feature descriptors and multiscale local binary patterns ( MLBP ). The authors were able to further improve the matching performance using race and gender information to reduce the target gallery size.
Q2. What is the proposed method for combining large feature size and small sample size?
In order to handle the combination of a large feature size and small sample size, an ensemble of linear discriminant classifiers called LFDA is proposed.
Q3. What other methods have been proposed to handle the SSS problem?
Other discriminant analysis methods have been proposed to handle the SSS problem, such as random sampling LDA [23], regularized LDA [24], and direct LDA [25].
Q4. Why are image descriptors not sufficiently verbose to describe a face?
Because most image descriptors are not sufficientlyverbose to fully describe a face image, the descriptors are computedover a set of uniformly distributed subregions of the face.
Q5. What is the main reason why the authors have not found a large number of forensic sketches?
the authors believe that with a larger number of forensic sketches, the authors could more properly train their discriminant and further improve the matching performance.
Q6. Why is the culprit being depicted in a forensic sketch considered a suspect?
This is because the culprit being depicted in a forensic sketch typically has committed a heinous crime (e.g., murder, rape, and armed robbery) that will receive a large amount of attention from investigators.
Q7. What is the approach to extract discriminant features?
A straightforward approach would be to apply classical subspace analysis (such as LDA) directly on , and to extract discriminant features for classification.
Q8. What are the key difficulties in matching forensic sketches?
The authors highlight two key difficulties in matching forensic sketches: 1) matching across image modalities and 2) performing face recognition despite possibly inaccurate depictions of the face.