Italy

We present a novel approach to developing an end-to-end learning framework to segment and classify major regions of a scientific document. We consider scientific document layout analysis as an object detection task over digital images, without any additional text features that need to be added into the network during the training process. Our technical objective is to implement transfer learning via fine-tuning of pre-trained networks and thereby demonstrate that this deep learning architecture is suitable for tasks that lack very large document corpora for training ab initio. As part of the experimental test bed for empirical evaluation of this approach, we created a merged multi-corpus data set for scientific publication layout detection tasks. Our results show good improvement with fine-tuning of a pre-trained base network using this merged data set, compared to the baseline convolutional neural network architecture.

ICPR 2020

Vision-Based Layout Detection from Scientific Literature using Recurrent Convolutional Neural Networks

document layout analysis

deep learning

information extraction

ICPR 2020 is the premier international conference on Pattern Recognition, gathering leading researchers from around the world to present their latest achievements in Pattern Analysis and Recognition, applied Artificial Intelligence and Machine Learning. The topic of the year is " Putting Artificial intelligence to work on Patterns". 
For the first time, ICPR has employed a two-round review process similar to journal submissions, with new papers allowed to be submitted in either the first or second round and papers submitted in the first round and not accepted allowed to be revised and re-submitted for second round review. 

In this edition of the **ICPR 2020** Conference, there are 1407 Oral/Poster presentations organized in five Tracks: 
- Artificial Intelligence, Machine Learning for Pattern Analysis; 
- Biometrics, Human Analysis and Behavior Understanding; Computer Vision, Robotics and Intelligent Systems; 
- Document and Media Analysis; and 
- Image and Signal Processing. 

The **ICPR 2020** Main Conference program also includes 7 Keynotes from prominent world-renowned scientists, 8 Tutorials, 28 Demos, 8 Challenges, 2 Panels, as well as an Industry/Sponsor Exhibition. In addition, 41 Workshops on a range of important current and emerging research topics complete the ICPR 2020 program. 


poster

Recently, various material recognition methods have
been introduced that uses a single color or light field camera. In
prior methods, color and texture information of an object are
used as key features. However, there exists fundamental limitation
in using color features for material recognition in that material
type can be characterized better by surface reflectance, visual
appearance rather than its color and textures. In this work, we
propose IR surface reflectance based material type recognition
method. We use off-the-shelf ToF camera to estimate the IR
reflectance of arbitrary surface. Material type recognition is
performed on both color and surface IR reflectance features. Several
network structures including gradual convolutional neural
network are proposed and verified for our material recognition
within our own 3D data sets.

Surface IR Reflectance Estimation and Material Recognition using ToF Camera

Although supervised person re-identification (RE-ID) has achieved great progress with deep learning, it requires time-consuming annotation of a large number of pedestrian identities. To reduce labeling cost, we attempt to reduce cross-camera identity annotations and exploit pedestrian attribute annotations as auxiliary information instead. The pedestrian attributes, such as outfit styles, contain coarse semantic knowledge. Although pedestrian attributes are annotated without exhaustive searching in a camera network, which is much easier than cross-camera identity annotation, ambiguity exists in attributes when different persons have similar outfits. To solve this problem, we propose an Attribute Similarity Guidance loss (ASG) to guide appearance feature learning for RE-ID by selective attribute similarity preservation to avoid the impact of such ambiguity. Finally, we develop an attribute-guided self training framework to jointly utilize attribute annotations, unlabeled data and limited labeled data for semi-supervised learning. Extensive experiments on Market-1501 and DukeMTMC-ReID show the superiority of our method for semi-supervised RE-ID.

Semi-supervised Person Re-identification by Attribute Similarity Guidance

We look at robustness in terms of reproducibility, which is challenging to improve and quantify in the context of dimensionality reduction for data visualization. We identify potential building blocks for achieving this kind of robustness and support these findings with a novel similarity measure for scatter plots. The building blocks are a restricted Boltzmann machine (RBM) pre-training for deep neural network (DNN) initilization and a deep autoencoder based regularization method.

Dimensionality Reduction for Data Visualization and Linear Classification - Robustness vs. Accuracy


This paper studies the problem of Person Re-Identification (ReID)
for large-scale applications. Recent research efforts have been devoted
to building complicated part models, which introduce considerably
high computational cost and memory consumption, inhibiting its practicability
in large-scale applications. This paper aims to develop a novel learning
strategy to find efficient feature embeddings while maintaining the
balance of accuracy and model complexity. More specifically, we find
by enhancing the classical triplet loss together with cross-entropy
loss, our method can explore the hard examples and build a discriminant
feature embedding yet compact enough for large-scale applications.
Our method is carried out progressively using Bayesian optimization,
and we call it the Progressive Learning Algorithm (PLA). Extensive
experiments on three large-scale datasets show that our PLA is comparable
or better than the state-of-the-arts. Especially, on the challenging
Market-1501 dataset, we achieve Rank-1=94.7%/mAP=89.4% while saving
at least 30\% parameters than strong part models.

Supervised Classification Using Graph-based Space Partitioning for Multiclass Problems

Significant advances have been achieved in face recognition in the last decade thanks to the development of deep learning methods. However, recognizing faces captured in uncontrolled environments is still a challenging problem for the scientific community. In these scenarios, the performance of most of existing deep learning based methods abruptly falls, due to the bad quality of the face images. In this work, we propose to use an activation map to represent the quality information in a face image. Different face regions are analyzed to determine their quality and then only those regions with good quality are used to perform the recognition using a given deep face model. For experimental evaluation, in order to simulate unconstrained environments, three challenging databases, with different variations in appearance, were selected: the Labeled Faces in the Wild Database, the Celebrities in Frontal-Profile in the Wild Database, and the AR Database. Three deep face models were used to evaluate the proposal on these databases and in all cases, the use of the proposed activation map allows the improvement of the recognition rates obtained by the original models in a range from 0.3 up to 31%. The obtained results experimentally demonstrated that the proposal is able to select those face areas with higher discriminative power and enough identifying information,

Quality-based Representation for Unconstrained Face Recognition

Unsupervised Sound Source Localization From Audio-Image Pairs Using Input Gradient Map

Infrared imaging has recently played an important role in a wide range of applications including video surveillance, robotics and night vision. However, infrared cameras often suffer from some limitations, essentially about low-contrast and blurred details. These problems contribute to the loss of observation of target objects in infrared images, which could limit the feasibility of different infrared imaging applications. In this paper, we mainly focus on the problem of pedestrian detection on thermal images. Particularly, we emphasis the need for enhancing the
visual quality of images before performing the detection step. To address that, we propose a novel thermal enhancement architecture called TE-GAN based on Generative Adversarial Network, and composed of two modules contrast enhancement and denoising with a post-processing step for edge restoration in
order to improve the overall image quality. The effectiveness of the proposed architecture is assessed by means of visual quality metrics and better results are obtained compared to the original thermal images and to the obtained results by other existing enhancement methods. These results have been conducted on a subset of KAIST dataset that we make available to encourage research in this direction. Using the same dataset, the impact of the proposed enhancement architecture has been demonstrated on the detection results by obtaining better performance with a significant margin using YOLOv3 detector.

Thermal Image Enhancement using Generative Adversarial Network for Pedestrian Detection

In this work, we propose an original CNN training strategy that brings together ideas from both dropout-like regularization methods and solutions that learn discriminative features. We propose a dropping criterion that, differently from dropout and its variants, is deterministic rather than random. It grounds on the empirical evidence that feature descriptors with larger L2-norm and highly-active nodes are strongly correlated to confident class predictions. Thus, our criterion guides towards dropping a percentage of the most active nodes of the descriptors, proportionally to the estimated class probability. We simultaneously train a per-sample scaling factor to balance the expected output across training and inference. This further allows us to keep high the descriptor's L2-norm, which we show enforces confident predictions. The combination of these two strategies resulted in our ``Probability Guided Maxout'' solution that acts as a training regularizer. We prove the above behaviors by reporting extensive image classification results on the CIFAR10, CIFAR100, and Caltech256 datasets.

Probability Guided Maxout

A novel automatic pose-invariant dental arch extraction and matching framework is developed for 3D dental identification using laser-scanned dental plasters. In our previous attempt [1-5], 3D point-based algorithms have been developed and they have shown a few advantages over existing 2D dental identifications. This study is a continuous effort in developing arch-based algorithms to extract and match dental arch feature in an automatic and pose-invariant way. As best as we know, this is the first attempt at automatic dental arch extraction and matching for 3D dental identification. A Radial Ray Algorithm (RRA) is proposed by projecting dental arch shape from 3D to 2D. This algorithm is fully automatic and fast. Preliminary identification result is obtained by matching 11 postmortem (PM) samples against 200 ante-mortem (AM) samples. 72.7% samples achieved top 5% accuracy. 90.9% samples achieved top 10% accuracy and all 11 samples (100%) achieved top 15.5% accuracy out of the 200-rank list. In addition, the time for identifying a single subject from 200 subjects has been significantly reduced from 45 minutes to 5 minutes by matching the extracted 2D dental arch. Although the extracted 2D arch feature is not as accurate and discriminative as the full 3D arch, it may serve as an important filter feature to improve the identification speed in future investigations.

3D Dental Biometrics: Automatic Pose-invariant Dental Arch Extraction and Matching

Finger vein recognition deals with the identification of subjects based on their venous pattern within the fingers. The recognition accuracy of finger vein recognition systems suffers from different internal and external factors. One of the major problems are misplacements of the finger during acquisition. In particular longitudinal finger rotation poses a severe problem for such recognition systems. The detection and correction of such rotations is a difficult task as typically finger vein scanners acquire only a single image from the vein pattern. Therefore, important information such as the shape of the finger or the depth of the veins within the finger, which are needed for the rotation detection, are not available. This work presents a CNN based rotation detector that is capable of estimating the rotational difference between vein images of the same finger without providing any additional information. The experiments executed not only show that the method delivers highly accurate results, but it also generalizes so that the trained CNN can also be applied on data sets which have not been included during the training of the CNN. Correcting the rotation difference between images using the CNN's rotation prediction leads to EER improvements between 50-260\% for a well-established vein-pattern based method (Maximum Curvature) on four public finger vein databases.

Premium content

Vision-Based Layout Detection from Scientific Literature using Recurrent Convolutional Neural Networks

Downloads

Next from ICPR 2020

Surface IR Reflectance Estimation and Material Recognition using ToF Camera

Similar lecture

Multiple Document Datasets Pre-training Improves Text Line Detection With Deep Neural Networks