arXiv subjects

Tianlong Ma

Publications and source records attributed to Tianlong Ma.

24 records · Page 2Linked to original sources

arXiv2021

Human-In-The-Loop Document Layout Analysis

Document layout analysis (DLA) aims to divide a document image into different types of regions. DLA plays an important role in the document content understanding and information extraction systems. Exploring a method that can use less data for effective training contributes to the development of DLA. We consider a Human-in-the-loop (HITL) collaborative intelligence in the DLA. Our approach was inspired by the fact that the HITL push the model to learn from the unknown problems by adding a small amount of data based on knowledge. The HITL select key samples by using confidence. However, using confidence to find key samples is not suitable for DLA tasks. We propose the Key Samples Selection (KSS) method to find key samples in high-level tasks (semantic segmentation) more accurately through agent collaboration, effectively reducing costs. Once selected, these key samples are passed to human beings for active labeling, then the model will be updated with the labeled samples. Hence, we revisited the learning system from reinforcement learning and designed a sample-based agent update strategy, which effectively improves the agent's ability to accept new samples. It achieves significant improvement results in two benchmarks (DSSE-200 (from 77.1% to 86.3%) and CS-150 (from 88.0% to 95.6%)) by using 10% of labeled data.

cs.CV↗

arXiv2021

Tutte polynomials of fan-like graphs with applications in benzenoid systems

We study the computation of the Tutte polynomials of fan-like graphs and obtain expressions of their Tutte polynomials via generating functions. As applications, Tutte polynomials, in particular, the number of spanning trees, of two kinds of benzenoid systems, i.e. pyrene chains and triphenylene chains, are obtained.

math.CO↗

arXiv2020

Edge-Aware Deep Image Deblurring

Image deblurring is a fundamental and challenging low-level vision problem. Previous vision research indicates that edge structure in natural scenes is one of the most important factors to estimate the abilities of human visual perception. In this paper, we resort to human visual demands of sharp edges and propose a two-phase edge-aware deep network to improve deep image deblurring. An edge detection convolutional subnet is designed in the first phase and a residual fully convolutional deblur subnet is then used for generating deblur results. The introduction of the edge-aware network enables our model with the specific capacity of enhancing images with sharp edges. We successfully apply our framework on standard benchmarks and promising results are achieved by our proposed deblur model.

cs.CV↗

arXiv2020

Scene Text Recognition with Temporal Convolutional Encoder

Texts from scene images typically consist of several characters and exhibit a characteristic sequence structure. Existing methods capture the structure with the sequence-to-sequence models by an encoder to have the visual representations and then a decoder to translate the features into the label sequence. In this paper, we study text recognition framework by considering the long-term temporal dependencies in the encoder stage. We demonstrate that the proposed Temporal Convolutional Encoder with increased sequential extents improves the accuracy of text recognition. We also study the impact of different attention modules in convolutional blocks for learning accurate text representations. We conduct comparisons on seven datasets and the experiments demonstrate the effectiveness of our proposed approach.

cs.CV↗

arXiv2020

Orbital-angular-momentum-based experimental test of Hardy's paradox for multisetting and multidimensional systems

Characterizing high-dimensional entangled states is of crucial importance in quantum information science and technology. Recent theoretical progress has been made to extend the Hardy's paradox into a general scenario with multisetting multidimensional systems, which can surpass the bound limited by the original version. Hitherto, no experimental verification has been conducted to verify such a Hardy's paradox, as most of previous experimental efforts were restricted to two-dimensional systems. Here, based on two-photon high-dimensional orbital angular momentum (OAM) entanglement, we report the first experiment to demonstrate the Hardy's paradox for multiple settings and multiple outcomes. We demonstrate the paradox for two-setting higher-dimensional OAM subspaces up to d = 7, which reveals that the nonlocal events increase with the dimension. Furthermore, we showcase the nonlocality with an experimentally recording probability of 36.77% for five-setting three-dimensional OAM subspace via entanglement concentration, and thus showing a sharper contradiction between quantum mechanics and classical theory.

physics.optics↗

arXiv2019

Fractional matching preclusion for generalized augmented cubes

The \emph{matching preclusion number} of a graph is the minimum number of edges whose deletion results in a graph that has neither perfect matchings nor almost perfect matchings. As a generalization, Liu and Liu recently introduced the concept of fractional matching preclusion number. The \emph{fractional matching preclusion number} of $G$ is the minimum number of edges whose deletion leaves the resulting graph without a fractional perfect matching. The \emph{fractional strong matching preclusion number} of $G$ is the minimum number of vertices and edges whose deletion leaves the resulting graph without a fractional perfect matching. In this paper, we obtain the fractional matching preclusion number and the fractional strong matching preclusion number for generalized augmented cubes. In addition, all the optimal fractional strong matching preclusion sets of these graphs are categorized.

math.CO↗