Search

This paper presents a method to improve the traditional mixture of deformable part models (MDPM) method from the learning perspective. First, an object part configuration learning algorithm based on group sparsity constraint is introduced to automatically discover the object part number, size, and location. The algorithm imposes two additional regularization terms in addition to the standard hinge loss function. The first term focuses on automatic part selection and the second term focuses on automatic part placement. Second, this paper introduces an improved MDPM training framework. The framework applies a learned transformation to normalize the prediction score from each individual deformable part model (DPM) into a pseudo probability such that the partition of the entire object appearance feature space becomes less sensitive to the prior distributions of different DPMs. Finally, the two proposed improvements are combined and formulated under the expectation-maximization framework. We evaluate our method mainly using the PASCAL VOC2007 and VOC2010 detection benchmarks and show that the proposed learning algorithms could increase the detection mean AP score by 2.4% and 0.9%, respectively, on these two data sets when using the proposed part selection method and the training algorithm. We also present further in-depth analysis of the proposed algorithm in the experiments.

Large-scale multimedia data generated in our daily life have increasingly motivated companies and researchers to discover valuable knowledge and hidden information by employing machine learning techniques. It is well known that this is an inherently time consuming job to deal with massive multimedia data locally. Thanks to the rapid development of cloud computing, the data owner is highly motivated to outsource these multimedia data, along with the computationally intensive processing tasks, to the cloud by leveraging its abundant resources for cost saving and flexibility. However, due to different trust domains, privacy issues also arise from the exposure to the cloud of private information that is intrinsically embedded in the outsourced data. As a key preprocessing step in machine learning, feature extractions have been extensively studied in the framework of privacy-preserving outsourced computation for their effectiveness in removing irrelevant and redundant data, and increasing learning accuracy. In this article, we provide a comprehensive survey of privacy-preserving outsourcing computation of various feature extraction algorithms in recent literature. Following a brief overview of each scheme, we present the primary technical hurdles needed to be addressed and then compare the state-of-theart solutions in terms of security and effectiveness. Then we discuss several promising future research directions with a focus on privacy-preserving machine learning in the cloud.

Advances in cloud computing have greatly motivated data owners to outsource their huge amount of personal multimedia data and/or computationally expensive tasks onto the cloud by leveraging its abundant resources for cost saving and flexibility. Despite the tremendous benefits, the outsourced multimedia data and its originated applications may reveal the data owner's private information, such as the personal identity, locations, or even financial profiles. This observation has recently aroused new research interest on privacy-preserving computations over outsourced multimedia data. In this paper, we propose an effective and practical privacy-preserving computation outsourcing protocol for the prevailing scale-invariant feature transform (SIFT) over massive encrypted image data. We first show that the previous solutions to this problem have either efficiency/security or practicality issues, and none can well preserve the important characteristics of the original SIFT in terms of distinctiveness and robustness. We then present a new scheme design that achieves efficiency and security requirements simultaneously with the preservation of its key characteristics, by randomly splitting the original image data, designing two novel efficient protocols for secure multiplication and comparison, and carefully distributing the feature extraction computations onto two independent cloud servers. We both carefully analyze and extensively evaluate the security and effectiveness of our design. The results show that our solution is practically secure, outperforms the state-of-the-art, and performs comparably with the original SIFT in terms of various characteristics, including rotation invariance, image scale invariance, robust matching across affine distortion, and addition of noise and change in 3D viewpoint and illumination.