We introduce a detection framework for dense crowd counting and eliminate the need prevalent density regression paradigm. Typical models predict an image as opposed to detecting every person. These methods, in general, fail localize persons accurate enough most applications other than counting. Hence, we adopt architecture that <italic xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">locate</i> s person crowd,...

Reconstructing a high-resolution 3D model of an object is challenging task in computer vision. Designing scalable and light-weight architectures crucial while addressing this problem. Existing point-cloud based reconstruction approaches directly predict the entire point cloud single stage. Although technique can handle low-resolution clouds, it not viable solution for generating dense, outputs. In work, we introduce DensePCR, deep pyramidal network that hierarchically predicts clouds...

Though Deep Neural Networks (DNN) show excellent performance across various computer vision tasks, several works their vulnerability to adversarial samples, i.e., image samples with imperceptible noise engineered manipulate the network's prediction. Adversarial sample generation methods range from simple complex optimization techniques. Majority of these generate adversaries through objectives that are tied pre-softmax or softmax output network. In this work we, (i) drawbacks such attacks,...

An unsupervised human action modeling framework can provide useful pose-sequence representation, which be utilized in a variety of pose analysis applications. In this work we propose novel temporal framework, embed the dynamics 3D human-skeleton joints to latent space an efficient manner. contrast end-to-end explored by previous works, disentangle task individual representation learning from actions as sequence embeddings. order realize continuous embedding manifold along with better...

We prove the #P-hardness of counting problems associated with various satisfiability, graph, and combinatorial problems, when restricted to planar instances. These include 3Sat, 1-3Sat, 1-Ex3Sat, Minimum Vertex Cover, Dominating Set, Feedback X3C, Partition Into Triangles, Clique Cover. also NP-completeness Ambiguous Satisfiability} [J. B. Saxe, Two Papers on Graph Embedding Problems, Tech. Report CMU-CS-80-102, Dept. Computer Science, Carnegie Mellon Univ., Pittsburgh, PA, 1980]...

Aiming towards human-level generalization, there is a need to explore adaptable representation learning methods with greater transferability. Most existing approaches independently address task-transferability and cross-domain adaptation, resulting in limited generalization. In this paper, we propose UM-Adapt - unified framework effectively perform unsupervised domain adaptation for spatially-structured prediction tasks, simultaneously maintaining balanced performance across individual tasks...

Freehand sketching is an inherently sequential process. Yet, most approaches for hand-drawn sketch recognition either ignore this aspect or exploit it in ad-hoc manner. In our work, we propose a recurrent neural network architecture object which exploits the long-term and structural regularities stroke data scalable Specifically, introduce Gated Recurrent Unit based framework leverages deep features weighted per-timestep loss to achieve state-of-the-art results on large database of freehand...

Estimation of 3D human pose from monocular image has gained considerable attention, as a key step to several human-centric applications. However, generalizability estimation models developed using supervision on large-scale in-studio datasets remains questionable, these often perform unsatisfactorily unseen in-the-wild environments. Though weakly-supervised have been proposed address this shortcoming, performance such relies availability paired some related task, 2D or multi-view pairs. In...

In this era of digital information explosion, an abundance data from numerous modalities is being generated as well archived everyday. However, most problems associated with training Deep Neural Networks still revolve around lack that rich enough for a given task. Data required not only initial model, but also future learning tasks such Model Compression and Incremental Learning. A diverse dataset may be used it feasible to store throughout the product life cycle due privacy issues or memory...

The ability to semantically interpret hand-drawn line sketches, although very challenging, can pave way for novel applications in multimedia. We propose SKETCHPARSE, the first deep-network architecture fully automatic parsing of freehand object sketches. SKETCHPARSE is configured as a two-level convolutional network. level contains shared layers common all categories. second number expert sub-networks. Each specializes sketches from categories which contain structurally similar parts....

Data provenance is essential for debugging query results, auditing data in cloud environments, and explaining outputs of Big analytics. A well-established technique to represent as annotations on instrument queries propagate these produce results annotated with provenance. However, even sophisticated optimizers are often incapable producing efficient execution plans instrumented queries, because their inherent complexity unusual structure. Thus, while instrumentation enables support...

Action recognition plays an important role in various applications, including smart homes and personal assistive robotics. In this paper, we propose algorithm for recognizing human actions using motion capture action data. Motion data provides accurate three dimensional positions of joints which constitute the skeleton. We model movement skeletal temporally order to classify action. The skeleton each frame sequence is represented as a 129 vector, component 3D angle made by joint with fixed...

Provenance for transactional updates is critical many applications such as auditing and debugging of transactions. Recently, we have introduced MV-semirings, an extension the semiring provenance model that supports Furthermore, proposed reenactment, a declarative form replay with capture, efficient non-invasive method computing this type provenance. However, approach limited to snapshot isolation (SI) concurrency control protocol while real world apply read committed version (RC-SI) improve...

We propose a novel recurrent network-based HDR deghosting method for fusing arbitrary length dynamic sequences. The proposed uses convolutional and architectures to generate visually pleasing, ghosting-free images. introduce new cell architecture, namely Self-Gated Memory (SGM) cell, that outperforms the standard LSTM while containing fewer parameters having faster running times. In SGM information flow through gate is controlled by multiplying gate's output function of itself. Additionally,...

We present a biologically inspired recurrent neural network (RNN) that efficiently detects changes in natural images. The model features sparse, topographic connectivity (st-RNN), closely modeled on the circuit architecture of “midbrain attention network.” deployed st-RNN challenging change blindness task, which must be detected discontinuous sequence Compared with conventional RNN, learned 9x faster and achieved state-of-the-art performance 15x fewer connections. An analysis low-dimensional...

As a form of visual representation, freehand line sketches are typically studied as an end product the sketching process. However, from recognition point view, one can also study various orderings and properties primitive strokes that compose sketch. Studying in this manner has enabled us to create novel sparse yet discriminative sketch-based representations for object categories which we term category-epitomes. Concurrently, epitome construction provides natural measure quantifying...

Despite the remarkable success of generative adversarial networks, their performance seems less impressive for diverse training sets, requiring learning discontinuous mapping functions. Though multi-mode prior or multi-generator models have been proposed to alleviate this problem, such approaches may fail depending on empirically chosen initial mode components. In contrast bottom-up approaches, we present GAN-Tree, which follows a hierarchical divisive strategy address multi-modal data....