arXiv subjects

Ling Shao

Publications and source records attributed to Ling Shao.

281 records · Page 16Linked to original sources

arXiv2017

From Zero-shot Learning to Conventional Supervised Classification: Unseen Visual Data Synthesis

Robust object recognition systems usually rely on powerful feature extraction mechanisms from a large number of real images. However, in many realistic applications, collecting sufficient images for ever-growing new classes is unattainable. In this paper, we propose a new Zero-shot learning (ZSL) framework that can synthesise visual features for unseen classes without acquiring real images. Using the proposed Unseen Visual Data Synthesis (UVDS) algorithm, semantic attributes are effectively utilised as an intermediate clue to synthesise unseen visual features at the training stage. Hereafter, ZSL recognition is converted into the conventional supervised problem, i.e. the synthesised visual features can be straightforwardly fed to typical classifiers such as SVM. On four benchmark datasets, we demonstrate the benefit of using synthesised unseen data. Extensive experimental results suggest that our proposed approach significantly improve the state-of-the-art results.

cs.CV↗

arXiv2017

Deep Sketch Hashing: Fast Free-hand Sketch-Based Image Retrieval

Free-hand sketch-based image retrieval (SBIR) is a specific cross-view retrieval task, in which queries are abstract and ambiguous sketches while the retrieval database is formed with natural images. Work in this area mainly focuses on extracting representative and shared features for sketches and natural images. However, these can neither cope well with the geometric distortion between sketches and images nor be feasible for large-scale SBIR due to the heavy continuous-valued distance computation. In this paper, we speed up SBIR by introducing a novel binary coding method, named \textbf{Deep Sketch Hashing} (DSH), where a semi-heterogeneous deep architecture is proposed and incorporated into an end-to-end binary coding framework. Specifically, three convolutional neural networks are utilized to encode free-hand sketches, natural images and, especially, the auxiliary sketch-tokens which are adopted as bridges to mitigate the sketch-image geometric distortion. The learned DSH codes can effectively capture the cross-view similarities as well as the intrinsic semantic correlations between different categories. To the best of our knowledge, DSH is the first hashing work specifically designed for category-level SBIR with an end-to-end deep architecture. The proposed DSH is comprehensively evaluated on two large-scale datasets of TU-Berlin Extension and Sketchy, and the experiments consistently show DSH's superior SBIR accuracies over several state-of-the-art methods, while achieving significantly reduced retrieval time and memory footprint.

cs.CV↗

arXiv2016

DAVE: A Unified Framework for Fast Vehicle Detection and Annotation

Vehicle detection and annotation for streaming video data with complex scenes is an interesting but challenging task for urban traffic surveillance. In this paper, we present a fast framework of Detection and Annotation for Vehicles (DAVE), which effectively combines vehicle detection and attributes annotation. DAVE consists of two convolutional neural networks (CNNs): a fast vehicle proposal network (FVPN) for vehicle-like objects extraction and an attributes learning network (ALN) aiming to verify each proposal and infer each vehicle's pose, color and type simultaneously. These two nets are jointly optimized so that abundant latent knowledge learned from the ALN can be exploited to guide FVPN training. Once the system is trained, it can achieve efficient vehicle detection and annotation for real-world traffic surveillance data. We evaluate DAVE on a new self-collected UTS dataset and the public PASCAL VOC2007 car and LISA 2010 datasets, with consistent improvements over existing algorithms.

cs.CV↗

arXiv2016

DAP3D-Net: Where, What and How Actions Occur in Videos?

Action parsing in videos with complex scenes is an interesting but challenging task in computer vision. In this paper, we propose a generic 3D convolutional neural network in a multi-task learning manner for effective Deep Action Parsing (DAP3D-Net) in videos. Particularly, in the training phase, action localization, classification and attributes learning can be jointly optimized on our appearancemotion data via DAP3D-Net. For an upcoming test video, we can describe each individual action in the video simultaneously as: Where the action occurs, What the action is and How the action is performed. To well demonstrate the effectiveness of the proposed DAP3D-Net, we also contribute a new Numerous-category Aligned Synthetic Action dataset, i.e., NASA, which consists of 200; 000 action clips of more than 300 categories and with 33 pre-defined action attributes in two hierarchical levels (i.e., low-level attributes of basic body part movements and high-level attributes related to action motion). We learn DAP3D-Net using the NASA dataset and then evaluate it on our collected Human Action Understanding (HAU) dataset. Experimental results show that our approach can accurately localize, categorize and describe multiple actions in realistic videos.

cs.CV↗

arXiv2015

Projection Bank: From High-dimensional Data to Medium-length Binary Codes

Recently, very high-dimensional feature representations, e.g., Fisher Vector, have achieved excellent performance for visual recognition and retrieval. However, these lengthy representations always cause extremely heavy computational and storage costs and even become unfeasible in some large-scale applications. A few existing techniques can transfer very high-dimensional data into binary codes, but they still require the reduced code length to be relatively long to maintain acceptable accuracies. To target a better balance between computational efficiency and accuracies, in this paper, we propose a novel embedding method called Binary Projection Bank (BPB), which can effectively reduce the very high-dimensional representations to medium-dimensional binary codes without sacrificing accuracies. Instead of using conventional single linear or bilinear projections, the proposed method learns a bank of small projections via the max-margin constraint to optimally preserve the intrinsic data similarity. We have systematically evaluated the proposed method on three datasets: Flickr 1M, ILSVR2010 and UCF101, showing competitive retrieval and recognition accuracies compared with state-of-the-art approaches, but with a significantly smaller memory footprint and lower coding complexity.

cs.CV↗

arXiv2015

Kernelized Multiview Projection

Conventional vision algorithms adopt a single type of feature or a simple concatenation of multiple features, which is always represented in a high-dimensional space. In this paper, we propose a novel unsupervised spectral embedding algorithm called Kernelized Multiview Projection (KMP) to better fuse and embed different feature representations. Computing the kernel matrices from different features/views, KMP can encode them with the corresponding weights to achieve a low-dimensional and semantically meaningful subspace where the distribution of each view is sufficiently smooth and discriminative. More crucially, KMP is linear for the reproducing kernel Hilbert space (RKHS) and solves the out-of-sample problem, which allows it to be competent for various practical applications. Extensive experiments on three popular image datasets demonstrate the effectiveness of our multiview embedding algorithm.

cs.CV↗

arXiv2014

Revised Annotations, Sex-Biased Expression, and Lineage-Specific Genes in the Drosophila melanogaster group

Here, we provide revised gene models for D. ananassae, D. yakuba, and D. simulans, which include UTRs and empirically verified intron-exon boundaries, as well as ortholog groups identified using a fuzzy reciprocal-best-hit blast comparison. Using these revised annotations, we perform differential expression testing using the cufflinks suite to provide a broad overview of differential expression between reproductive tissues and the carcass. We identify thousands of genes that are differentially expressed across tissues in D. yakuba and D. simulans, with roughly 60% agreement in expression patterns of orthologs in D. yakuba and D. simulans. We identify several cases of putative polycistronic transcripts, pointing to a combination of transcriptional read-through in the genome as well as putative gene fusion and fission events across taxa. We furthermore identify hundreds of lineage specific genes in each species with no blast hits among transcripts of any other Drosophila species, which are candidates for neofunctionalized proteins and a potential source of genetic novelty.

q-bio.GN↗

arXiv2014

Tandem duplications and the limits of natural selection in Drosophila yakuba and Drosophila simulans

Tandem duplications are an essential source of genetic novelty, and their variation in natural populations is expected to influence adaptive walks. Here, we describe evolutionary impacts of recently-derived, segregating tandem duplications in Drosophila yakuba and Drosophila simulans. We observe an excess of duplicated genes involved in defense against pathogens, insecticide resistance, chorion development, cuticular peptides, and lipases or endopeptidases associated with the accessory glands, suggesting that duplications function in Red Queen dynamics and rapid evolution. We document evidence of widespread selection on the D. simulans X, suggesting adaptation through duplication is common on the X. Despite the evidence for positive selection, duplicates display an excess of low frequency variants consistent with largely detrimental impacts, limiting the variation that can effectively facilitate adaptation. Although we observe hundreds of gene duplications, we show that segregating variation is insufficient to provide duplicate copies of the entire genome, and the number of duplications in the population spans 13.4\% of major chromosome arms in D. yakuba and 9.7\% in D. simulans. Whole gene duplication rates are low at $1.17\times10^{-9}$ per gene per generation in D. yakuba and $6.03\times10^{-10}$ per gene per generation in D. simulans, suggesting long wait times for new mutations on the order of thousands of years for the establishment of sweeps. Hence, in cases where adaption depends on individual tandem duplications, evolution will be severely limited by mutation. We observe low levels of parallel recruitment of the same duplicated gene in different species, suggesting that the span of standing variation will define evolutionary outcomes in spite of convergence across gene ontologies consistent with rapidly evolving phenotypes.} }

q-bio.PE↗

arXiv2014

Landscape of standing variation for tandem duplications in Drosophila yakuba and Drosophila simulans

We have used whole genome paired-end Illumina sequence data to identify tandem duplications in 20 isofemale lines of D. yakuba, and 20 isofemale lines of D. simulans and performed genome wide validation with PacBio long molecule sequencing. We identify 1,415 tandem duplications that are segregating in D. yakuba as well as 975 duplications in D. simulans, indicating greater variation in D. yakuba. Additionally, we observe high rates of secondary deletions at duplicated sites, with 8% of duplicated sites in D. simulans and 17% of sites in D. yakuba modified with deletions. These secondary deletions are consistent with the action of the large loop mismatch repair system acting to remove polymorphic tandem duplication, resulting in rapid dynamics of gain and loss in duplicated alleles and a richer substrate of genetic novelty than has been previously reported. Most duplications are present in only single strains, suggesting deleterious impacts are common. D. simulans shows larger numbers of whole gene duplications in comparison to larger proportions of gene fragments in D. yakuba. D. simulans displays an excess of high frequency variants on the X chromosome, consistent with adaptive evolution through duplications on the D. simulans X or demographic forces driving duplicates to high frequency. We identify 78 chimeric genes in D. yakuba and 38 chimeric genes in D. simulans, as well as 143 cases of recruited non-coding sequence in D. yakuba and 96 in D. simulans, in agreement with rates of chimeric gene origination in D. melanogaster. Together, these results suggest that tandem duplications often result in complex variation beyond whole gene duplications that offers a rich substrate of standing variation that is likely to contribute both to detrimental phenotypes and disease, as well as to adaptive evolutionary change.

q-bio.GN↗

arXiv2011

An Algorithm for Repairing Low-Quality Video Enhancement Techniques Based on Trained Filter

Multifarious image enhancement algorithms have been used in different applications. Still, some algorithms or modules are imperfect for practical use. When the image enhancement modules have been fixed or combined by a series of algorithms, we need to repair them as a whole part without changing the inside. This report aims to find an algorithm based on trained filters to repair low-quality image enhancement modules. A brief review on basic image enhancement techniques and pixel classification methods will be presented, and the procedure of trained filters will be described step by step. The experiments and result comparisons for this algorithm will be described in detail.

cs.CV↗

arXiv2010

Dynamic Feature Description in Human Action Recognition

This work aims to present novel description methods for human action recognition. Generally, a video sequence can be represented as a collection of spatial temporal words by detecting space-time interest points and describing the unique features around the detected points (Bag of Words representation). Interest points as well as the cuboids around them are considered informative for feature description in terms of both the structural distribution of interest points and the information content inside the cuboids. Our proposed description approaches are based on this idea and making the feature descriptors more discriminative.

cs.HC↗