Italy


Handwritten Text Recognition (HTR) in free-layout pages is a valuable yet challenging task which aims to automatically understand handwritten texts. State-of-the-art approaches in this field usually encode input images with Convolutional Neural Networks, whose kernels are typically defined on a fixed grid and focus on all input pixels independently. However, this is in contrast with the sparse nature of handwritten pages, in which only pixels representing the ink of the writing are useful for the recognition task. Furthermore, the standard convolution operator is not explicitly designed to take into account the great variability in shape, scale, and orientation of handwritten characters. To overcome these limitations, we investigate the use of deformable convolutions for handwriting recognition. The kernel of this type of convolution deforms according to the content of the neighborhood, and can therefore be more adaptable to geometric variations and other deformations of the text. Experiments conducted on the IAM and RIMES datasets demonstrate that the use of deformable convolutions is a promising direction for the design of novel architectures for handwritten text recognition.

ICPR 2020

Watch Your Strokes: Improving Handwritten TextRecognition with Deformable Convolutions

deformable convolutions

document analysis

handwritten text recognition

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

Many computer vision applications require robust and efficient estimation of camera geometry. The robust estimation is usually based on solving camera geometry problems from a minimal number of input data measurements, i.e., solving minimal problems, in a RANSAC-style framework. Minimal problems often result in complex systems of polynomial equations. The existing state-of-the-art methods for solving such systems are either based on Grobner bases and the action matrix method, which have been extensively studied and optimised in the recent years or recently proposed approach based on a sparse resultant computation using an extra variable.

In this paper, we study an interesting alternative sparse resultant-based method for solving sparse systems of polynomial equations by hiding one variable. This approach results in a larger eigenvalue problem than the action matrix and extra variable sparse resultant-based methods; however, it does not need to compute an inverse or elimination of large matrices that may be numerically unstable. The proposed approach includes several improvements to the standard sparse resultant algorithms, which significantly improves the efficiency and stability of the hidden variable resultant-based solvers as we demonstrate on several interesting computer vision problems. We show that for the studied problems, our sparse resultant based approach leads to more stable solvers than the state-of-the-art Grobner basis as well as existing sparse resultant-based solvers, especially in close to critical configurations. Our new method can be fully automated and incorporated into existing tools for the automatic generation of efficient minimal solvers.

Computing stable resultant based minimal solvers by hiding a variable

Activity recognition in children and infants is important in applications such as safety monitoring, behavior assessment, and child-robot interaction, among others. However, it differs from activity recognition in adults not only because body poses and proportions are different, but also because of the way in which actions are performed. This paper addresses the problem of infant action classification in challenging conditions. The actions are performed in a pediatric rehabilitation environment in which not only infants but also robots and adults are present, with the infant being one of the smallest actors in the scene. We propose a multiview action classification system based on Faster R-CNN and LSTM networks, which fuses information from different views by using learnable fusion coefficients derived from detection confidence scores. The proposed system is view-independent, learns features that are close to view-invariant, and can handle new or missing views at test time. Our approach outperforms the state-of-the-art baseline model for a small dataset (2 subjects, 10-24 months old) by 11.4% in terms of average classification accuracy in four classes (crawl, sit, stand and walk). Moreover, experiments in an extended dataset (6 subjects, 8-24 months old) show that the proposed fusion strategy outperforms all the alternative fusion methods studied.



A Detection-based Approach to Multiview Action Classification in Infants

This work aimed at giving a comprehensive and indetailed guide on the route to fine-tuning Convolutional Neural Networks (CNNs) for glaucoma screening. Transfer learning consists in a promising alternative to train CNNs from scratch, to avoid the huge data and resources requirements. After a thorough study of five state-of-the-art CNNs architectures, a complete and well-explained strategy for fine-tuning these networks is proposed, using hyperparameter grid-searching and two-phase training approach. Excellent performance is reached on model evaluation, with a 0.9772 AUROC validation rate, giving arise to reliable glaucoma diagosis-help systems. Also, a benchmark analysis is conducted across all fine-tuned models, studying them according to performance indices such as model complexity and size, AUROC density and inference time. This in-depth analysis allows a rigorous comparison between model characteristics, and is useful for giving practioners important trademarks for prospective applications and deployments.

Fine-tuning Convolutional Neural Networks: a comprehensive guide and benchmark analysis for glaucoma screening

Synthetic image generation provides the ability to efficiently produce large quantities of labeled data, which addresses both the data volume requirements of state-of-the-art vision systems and the expense of manually labeling data. However, systems trained on synthetic data typically under-perform systems trained on realistic data due to mismatch between the synthetic and realistic data distributions. Domain Randomization (DR) is a method of broadening a synthetic data distribution to encompass a realistic data distribution and provides better performance when the exact characteristics of the realistic data distribution are not known or cannot be simulated. However, there is no consensus in the literature on the best method of performing DR. We propose a novel method of ranking DR methods by directly measuring the difference between realistic and DR data distributions. This avoids the need to measure task-specific performance and the associated expense of training and evaluation. We compare different methods for measuring distribution differences, including the Wasserstein and Fréchet Inception distances. We also examine the effect of performing this evaluation directly on images and features generated by an image classification backbone. Finally, we show that the ranking generated by our method is reflected in actual task performance.

Quantifying the Use of Domain Randomization

Knowledge distillation (KD) is commonly deemed as an effective model compression technique in which a compact model (student) is trained under the supervision of a larger pretrained model or an ensemble of models (teacher). Various techniques have been proposed since the original formulation, which mimics different aspects of the teacher such as the representation space, decision boundary, or intra-data relationship. Some methods replace the one-way knowledge distillation from a static teacher with collaborative learning between a cohort of students. Despite the recent advances, a clear understanding of where knowledge resides in a deep neural network and an optimal method for capturing knowledge from teacher and transferring it to student remains an open question. In this study, we provide an extensive study on nine different KD methods which covers a broad spectrum of approaches to capture and transfer knowledge. We demonstrate the versatility of the KD framework on different datasets and network architectures under varying capacity gaps between the teacher and student. The study provides intuition for the effects of mimicking different aspects of the teacher and derives insights from the performance of the different distillation approaches to guide the design of more effective KD methods. Furthermore, our study shows the effectiveness of the KD framework in learning efficiently under varying severity levels of label noise and class imbalance, consistently providing generalization gains over standard training. We emphasize that the efficacy of KD goes much beyond a model compression technique and it should be considered as a general-purpose training paradigm which offers more robustness to common challenges in the real-world datasets compared to the standard training procedure.

Knowledge Distillation Beyond Model Compression


Objective: In the last years, Robot-assisted partial nephrectomy (RAPN) is establishing as elected treatment for renal cell carcinoma (RCC). Reduced field of view, field occlusions by surgical tools, and reduced maneuverability may potentially cause accidents, such as unwanted vessel resection with conse- quent bleeding. Surgical Data Science (SDS) can provide effective context-aware tools for supporting surgeons. However, currently no tools have been exploited for automatic vessels segmentation from nephrectomy laparoscopic videos. Herein, we propose a new approach based on adversarial Fully Convolutional Neural Net- works (FCNNs) to kidney vessel segmentation from nephrectomy laparoscopic vision. Methods: The proposed approach enhances existing segmentation framework by (i) encoding 3D kernels for spatio-temporal features extraction to enforce pixel connectivity in time, and (ii) perform training in adversarial fashion, which constrains vessels shape. Results: We performed a preliminary study using 8 different RAPN videos (1871 frames), the first in the field, achieving a median Dice Similarity Coefficient of 71.76%. Conclusions: Results showed that the proposed approach could be a valuable solution with a view to assist surgeon during RAPN.

NephCNN: A Deep-Learning Framework for Vessel Segmentation in Nephrectomy Laparoscopic Videos


One of the main factors that contributed to the large advances in autonomous driving is the advent of deep learning. For safer self-driving vehicles, one of the problems that has yet to be solved completely is lane detection. Since methods for this task have to work in real-time (+30 FPS), they not only have to be effective (i.e., have high accuracy) but they also have to be efficient (i.e., fast). In this work, we present a novel method for lane detection that uses as input an image from a forward-looking camera mounted in the vehicle and outputs polynomials representing each lane marking in the image, via deep polynomial regression. The proposed method is shown to be competitive with existing state-of-the-art methods in the TuSimple dataset while maintaining its efficiency (115 FPS). Additionally, extensive qualitative results on two additional public datasets are presented, alongside with limitations in the evaluation metrics used by recent works for lane detection. Finally, we provide source code and trained models that allow others to replicate all the results shown in this paper, which is surprisingly rare in state-of-the-art lane detection methods.

PolyLaneNet: Lane Estimation via Deep Polynomial Regression

In the recognition problem, a canonical form that
expresses the spatio-temporal relation of concepts for a given
class can potentially increase accuracy. Concepts are defined as
attributes that can be recognized using a soft matching paradigm.
We consider the specific case study of American Sign Language
(ASL) to show that canonical forms of classes can be used to
recognize unseen gestures. There are several advantages of a
canonical form of gestures including translation between gestures,
gesture-based searching, and automated transcription of gestures
into any spoken language. We applied our technique to two
independently collected datasets: a) IMPACT Lab dataset: 23
ASL gestures each executed three times from 130 first time ASL
learners as training data and b) ASLTEXT dataset: 190 gestures
each executed six times on an average. Our technique was able to
recognize 19 arbitrarily chosen previously unseen gestures in the
IMPACT dataset from seven individuals who are not a part of 130
and 34 unseen gestures from the ASLTEXT dataset without any
retraining. Our normalized accuracy on the ASLTEXT dataset
is 66% which is 13.6 % higher than the state-of-art technique.

Concept Embedding through Canonical Forms: A Case Study on Zero-Shot American Sign Language Recognition

In this paper, we propose the use of Optical Coherence
Tomography (OCT) imaging for the problem of iris
presentation attack (PA) detection. We assess its viability by
comparing its performance with respect to traditional iris imaging
modalities, viz., near-infrared (NIR) and visible spectrum.
OCT imaging provides a cross-sectional view of an eye, whereas
traditional imaging provides 2D iris textural information. PA
detection is performed using three state-of-the-art deep architectures
(VGG19, ResNet50 and DenseNet121) to differentiate
between bonafide and PA samples for each of the three imaging
modalities. Experiments are performed on a dataset of 2,169
bonafide, 177 Van Dyke eyes and 360 cosmetic contact images
acquired using all three imaging modalities under intra-attack
(known PAs) and cross-attack (unknown PAs) scenarios. We
observe promising results demonstrating OCT as a viable solution
for iris presentation attack detection.

Viability of Optical Coherence Tomography for Iris Presentation Attack Detection


In this work, we tackle the problem of estimating a camera capability to preserve fine texture details at a given lighting condition. Importantly, our texture preservation measurement should coincide with human perception. Consequently, we formulate our problem as a regression one and we introduce a deep convolutional network to estimate texture quality score. At training time, we use ground-truth quality scores provided by expert human annotators in order to obtain a subjective quality measure. In addition, we propose a region selection method to identify the image regions that are better suited at measuring perceptual quality. Finally, our experimental evaluation shows that our learning-based approach outperforms existing methods and that our region selection algorithm consistently improves the quality estimation.

Downloads

Next from ICPR 2020

Computing stable resultant based minimal solvers by hiding a variable

Similar lecture

A Few-shot Learning Approach for Historical Ciphered Manuscript Recognition