Person re-identification by symmetry-driven accumulation of local features


In this paper, we present an appearance-based method for person re-identification. It consists in the extraction of features that model three complementary aspects of the human appearance: the overall chromatic content, the spatial arrangement of colors into stable regions, and the presence of recurrent local motifs with high entropy. All this information is derived from different body parts, and weighted opportunely by exploiting symmetry and asymmetry perceptual principles. In this way, robustness against very low resolution, occlusions and pose, viewpoint and illumination changes is achieved. The approach applies to situations where the number of candidates varies continuously, considering single images or bunch of frames for each individual. It has been tested on several public benchmark datasets (ViPER, iLIDS, ETHZ), gaining new state-of-the-art performances.

Download full-text


Available from: Loris Bazzani,
  • Source
    • "Specifically, in person Re-ID benchmarks the head is always located at the top of images, torso in the middle, and legs at the bottom. This typical structure has been exploited in designing discriminative features [10] "
    [Show abstract] [Hide abstract]
    ABSTRACT: The group membership prediction (GMP) problem involves predicting whether or not a collection of instances share a certain semantic property. For instance, in kinship verification given a collection of images, the goal is to predict whether or not they share a {\it familial} relationship. In this context we propose a novel probability model and introduce latent {\em view-specific} and {\em view-shared} random variables to jointly account for the view-specific appearance and cross-view similarities among data instances. Our model posits that data from each view is independent conditioned on the shared variables. This postulate leads to a parametric probability model that decomposes group membership likelihood into a tensor product of data-independent parameters and data-dependent factors. We propose learning the data-independent parameters in a discriminative way with bilinear classifiers, and test our prediction algorithm on challenging visual recognition tasks such as multi-camera person re-identification and kinship verification. On most benchmark datasets, our method can significantly outperform the current state-of-the-art.
  • Source
    • "Appearance of a person is the visible foreground image after background subtraction [2]. Similarly, appearance-based methods rely on clothes, visual parts or perceptual principles to extract features for object recognition [2] [3]. Texture on the other hand, contains structural arrangement of a surface and its relation with the environments [4]. "
    [Show abstract] [Hide abstract]
    ABSTRACT: This paper presents an automatic, machine vision based, military personnel identification and classification system. Classification was done using a Support Vector Machine (SVM) on sets of Army, Air Force and Navy camouflage uniform personnel datasets retrieved from google images and selected military websites. In the proposed system, the arm of service of personnel is recognised by the camouflage and the type of cap badge/logo of a persons uniform. The detailed analysis done include; camouflage and plain cap differentiation using Gray Level Co- occurrence Matrix (GLCM) texture features; Army, Air Force and Navy camouflaged uniforms differentiation using GLCM texture and colour histogram bin features; plain cap differentiation using Speed Up Robust Feature (SURF) on the cap badge. Correlation-based Feature Selection (CFS) was used to improve recognition by selecting discriminating features, thereby speeding the classification process. With this method success rates recorded during the analysis include 94% for camouflage appearance category, 100%, 90% and 100% rates of plain and camouflage cap categories for Army, Air Force and Navy respectively. Similarly, using SURF features on the cap badge in the top region of the segmented human part of top and bottom; the plain cap badge of the military personnel was accurately categorised. By this, we have shown that the proposed method can be integrated into a face recognition system, which recognises an individual and determine the arm of service the person belongs. Such a system can be used to enhance the security of a military base or facility. Substantial analysis has been carried out and results after comparison with two other techniques prove that the proposed method can correctly classify military personnel into various arms of service. Accurate recognition was recorded with the proposed technique.
  • Source
    • "This category mainly focuses on pursing discriminative image representation robust to view, illumination, pose, background variations. In [10], Farenzena et. al augmented maximally stable color regions with histograms and recurrent local color patches. "
    [Show abstract] [Hide abstract]
    ABSTRACT: Person re-identification aims to match people across non-overlapping camera views, which is an important but challenging task in video surveillance. In order to obtain a robust metric for matching, metric learning has been introduced recently. Most existing works focus on seeking a Mahalanobis distance by employing sparse pairwise constraints, which utilize image pairs with the same person identity as positive samples, and select a small portion of those with different identities as negative samples. However, this training strategy has abandoned a large amount of discriminative information, and ignored the relative similarities. In this paper, we propose a novel Relevance Metric Learning method with Listwise Constraints (RMLLC) by adopting listwise similarities, which consist of the similarity list of each image with respect to all remaining images. By virtue of listwise similarities, RMLLC could capture all pairwise similarities, and consequently learn a more discriminative metric by enforcing the metric to conserve predefined similarity lists in a low dimensional projection subspace. Despite the performance enhancement, RMLLC using predefined similarity lists fails to capture the relative relevance information, which is often unavailable in practice. To address this problem, we further introduce a rectification term to automatically exploit the relative similarities, and develop an efficient alternating iterative algorithm to jointly learn the optimal metric and the rectification term. Extensive experiments on four publicly available benchmarking datasets are carried out and demonstrate that the proposed method is significantly superior to state-of-the-art approaches. The results also show that the introduction of the rectification term could further boost the performance of RMLLC.
    IEEE Transactions on Image Processing 08/2015; 24(12). DOI:10.1109/TIP.2015.2466117 · 3.63 Impact Factor
Show more