Brazil

In this paper, we proposed and evaluated the adoption of a capsule network-based (CapsNet-based) model rather than the convolutional neural network-based (CNN-based) models which are predominant in crowd counting tasks. The aim is to join the task of generating a high-quality density map from a single image along with producing a more precise estimate of the number of people. CapsNet-based model has a strong capacity of representation and a powerful dynamic routing mechanism that could address the drawback of a limited number of training samples. The replacement of the scalar values of CNN by vectors when using a CapsNet allows learning more discriminative features, which contributes to generating high-quality density maps, and thus a more precise number of individuals in crowd scenes. Experimental results show that our proposal presents competitive results concerning state-of-the-art, but with a 59.2% reduction in the number of parameters.

IWSSIP 2020

Dense Crowd Counting with Capsule Networks

density map

convolutional neural networks (cnn)

crowd density estimation

capsule network (capsnet)

crowd counting

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. 


ICPR 2020

technical paper

The 27th International Conference on Systems, Signals and Image Processing.

Acute leukemia is a cancer-related to a bone marrow abnormality. It is more common in children and young adults. This type of leukemia generates unusual cell growth in a short period, requiring a quick start of treatment. Acute Lymphoid Leukemia (ALL) and Acute Myeloid Leukemia (AML) are the main responsible for deaths caused by this cancer. The classification of these two leukemia types on blood slide images is a vital process of and automatic system that can assist doctors in the selection of appropriate treatment. This work presents a convolutional neural networks (CNNs) architecture capable of differentiating blood slides with ALL, AML and Healthy Blood Slides (HBS). The experiments were performed using 16 datasets with 2,415 images, and the accuracy of 97.18% and a precision of 97.23% were achieved. The proposed model results were compared with the results obtained by the state of the art methods, including also based on CNNs.

Convolution Neural Network Models for Acute Leukemia Diagnosis

The time required for a vehicle to travel different routes in the daily traffic of large cities varies and changes constantly, impacting the daily lives of everyone dwelling in those cities. Trying to predict such time is essential to evolve in understanding the behavior of vehicle traffic. On the other hand, due to the vast amount of data generated in this context, it is necessary to use new ways of dealing with this problem. This paper presents an exploratory analysis of the behavior of batch and stream learning algorithms for predicting the trip duration time for vehicles going through different routes. We understand that batch learning algorithms are not necessarily adequate for being used in stream mining situations. However, we would like to have a testbed to analyze the behavior of stream learning algorithms. For our experimental analysis, we used real data from three specific routes. The results show that the use of data stream learning for this domain yields promising results

An Exploratory Study Using Stream Learning Algorithms to Predict Duration Time of Vehicle Routes

The prediction of secondary protein structures is one of the classic problems of bioinformatics and has several practical applications.In this work, we present an ensemble of bidirectional recurrent networks with random forests. The fusion was performed using weights for each of the classifiers and classes, found by searching through genetic algorithms. To evaluate our method, we used the PDB, CB6133 and CB513 datasets. We achieved 73.1%, 59.1% and 55.8% Q8 accuracy on PDB, CB6133 and CB513 proteins, respectively, whereas 81.5% Q3 accuracy on PDB proteins using only amino acid sequence information. In order to compare our results against the literature, we also evaluated our method using amino acid sequence and sequence profile features, achieving 73.4% and 68.9% Q8 accuracy on CB6133 and CB513. Our method yielded good results when using only the amino acid sequence and presents competitive results compared tothe literature when using amino acid sequence information and protein sequence similarity.

Ensemble of Bidirectional Recurrent Networks and Random Forests for Protein Secondary Structure Prediction

To make possible the exploration of oil and natural gas in maritime waters, it is necessary to develop high technology to support the projects of drilling and exploration of submarine wells. One of the equipment that enables offshore exploration is the flexible pipes. Considering the high cost and complexity of the flexible pipe manufacturing process, the objective of this work is to analyze the effect of the occurrence of whitening stress identified as bleaching observed in the polymeric material known as Polyfluoride of Vinylidene (PVDF), which is used in the manufacture of inner liner layers in flexible pipes. The proposed methodology is to develop a prediction model based on classification, which evaluates the whitening bleaching by extracting knowledge from an image database. Through the proposed methodology, the most relevant parameters associated with the study problem that can define classification rules areidentified. As a result, some relationships between some characters, fruits of image data extraction and product characteristics were identified through decision trees. The conclusion is that these relationships provides support for decision making when peformed by a domain expert.

Analysis of Non-Conformity in Flexible Pipe through Data Mining from Image Bank

The Neutrinos Angra Experiment aims at monitoring nuclear plants by detecting antineutrino particles coming out from its fuel burn-up process. The Neutrinos Angra Collaboration developed a fully-equipped detector to accomplish this task. It
is currently in operation on the surface and next to the dome of the Angra II nuclear reactor. Selecting antineutrinos events on the surface is a challenge task due to the high level of background noise produced by cosmic ray particles. One of the
main parameters used to select antineutrino events is the number of photons acquired by the detector’s sensors. This quantity is estimated based on the signals generated at the output of the readout electronics. If any of those signals is corrupted, the estimation fails, compromising the Experiment performance. This
work proposes a study of the performance of some classifier algorithms applied to identify corrupted signals for the Neutrinos Angra Experiment. If corrupted signals can be identified, they could go through a recovering process in order to improve the Experiment’s events-selection algorithms.

Identification and Classification of Corrupted Signals for the Neutrinos Angra Experiment

We discuss the use of Associative Rule Learning (ARL) in the interface design of a 3D interactive visualization of temporal medical data. Our results indicate that ARL can be a valuable resource to improve the design of Information Visualization interface tools as well as to customize visualizations according to the user´s preferences.

Machine Learning in Information Visualization: Using Rule-based Learning Algorithms to Personalize Interfaces

The Neutrinos Angra Experiment aims at developing a system capable of monitoring the fuel-burning process of nuclear reactors. This system is based on an antineutrino detector located a few meters far from the core of the Angra II nuclear reactor. One of the main parameters used to select antineutrino events is the released energy left by a particle when interacting with the detector medium which, in turn, is proportional to the amplitude and area of the signals generated by the detector’s transducers. However, the estimation of such quantities may lose accuracy
due to saturation effects caused by the experiment electronics, making it necessary to apply methods to compensate or correct such effects. This work proposes to implement and test machine learning algorithms designed to estimate the energy of saturated signals.

Amplitude Estimators Applied to Saturated Signals for the v-Angra Experiment

This work presents a technique of working together to accumulate learning between various artificial neural networks, the technique has been successfully adapted to a convolutional network and has shown promising results in the context of traffic light classification.

Optimizing CNN's using Cumulative Learning with Auxiliary Networks for Traffic Light Recognition

This presentation shows a study about the feasibility of combining sustained vowels for computer-based pathological voice characterization. Although recent work has shown that single sustained vowels enable accurate voice characterization, the literature lacks evidence about combining them. Here we propose a pipeline for pathological voice characterization and classification that encompasses a feature extraction method based on the Haar wavelet. We also evaluate the benefits of our pipeline when using an input signal that comprises three different sustained vowels instead of single ones. We conducted experiments on samples of sustained vowels /a/, /i/, and /u/ from two public datasets, exploring the wavelet decomposition levels in the range of 4 to 18.

Combined Sustained Vowels Improve the Performance of the Haar Wavelet for Pathological Voice Characterization

The skin is an essential system for the human body, also the most susceptible to diseases. The diagnosis of these diseases is made mainly through dermatoscopy images, a noninvasive procedure. This paper presents a semi-automatic method that can assist the physician in monitoring the skin lesion’s evolution. Our approach applies the SLIC0 algorithm for superpixels generation, and texture features extraction. Subsequently, the algorithm Seeded Fuzzy C-means clusters the superpixels based on marked regions made by a specialist. Finally, the images undergo a post-processing step, eliminating noise and smoothing edges. We used 1380 images from four public datasets: the PH2 and DermIs to settle the method parameters, and ISIC 2016 and ISIC 2017 to the performance evaluation. Our tests show that by marking 1% of the superpixels, our method achieves an accuracy of 96.16% in ISIC 2016 and 95.34% in ISIC 2017. According to the literature, the segmentation was considered excellent in 93.49% of the images in ISIC 2016 and on 79.44% of ISIC 2017 images.

Premium content

Downloads

Next from IWSSIP 2020

Convolution Neural Network Models for Acute Leukemia Diagnosis

Similar lecture

Learning from Web Data: Improving Crowd Counting via Semi-Supervised Learning