-
Analysis of Social Robotic Navigation approaches: CNN Encoder and Incremental Learning as an alternative to Deep Reinforcement Learning
Authors:
Janderson Ferreira,
Agostinho A. F. Júnior,
Letícia Castro,
Yves M. Galvão,
Pablo Barros,
Bruno J. T. Fernandes
Abstract:
Dealing with social tasks in robotic scenarios is difficult, as having humans in the learning loop is incompatible with most of the state-of-the-art machine learning algorithms. This is the case when exploring Incremental learning models, in particular the ones involving reinforcement learning. In this work, we discuss this problem and possible solutions by analysing a previous study on adaptive c…
▽ More
Dealing with social tasks in robotic scenarios is difficult, as having humans in the learning loop is incompatible with most of the state-of-the-art machine learning algorithms. This is the case when exploring Incremental learning models, in particular the ones involving reinforcement learning. In this work, we discuss this problem and possible solutions by analysing a previous study on adaptive convolutional encoders for a social navigation task.
△ Less
Submitted 5 September, 2020; v1 submitted 18 August, 2020;
originally announced August 2020.
-
Performance Improvement of Path Planning algorithms with Deep Learning Encoder Model
Authors:
Janderson Ferreira,
Agostinho A. F. Júnior,
Yves M. Galvão,
Pablo Barros,
Sergio Murilo Maciel Fernandes,
Bruno J. T. Fernandes
Abstract:
Currently, path planning algorithms are used in many daily tasks. They are relevant to find the best route in traffic and make autonomous robots able to navigate. The use of path planning presents some issues in large and dynamic environments. Large environments make these algorithms spend much time finding the shortest path. On the other hand, dynamic environments request a new execution of the a…
▽ More
Currently, path planning algorithms are used in many daily tasks. They are relevant to find the best route in traffic and make autonomous robots able to navigate. The use of path planning presents some issues in large and dynamic environments. Large environments make these algorithms spend much time finding the shortest path. On the other hand, dynamic environments request a new execution of the algorithm each time a change occurs in the environment, and it increases the execution time. The dimensionality reduction appears as a solution to this problem, which in this context means removing useless paths present in those environments. Most of the algorithms that reduce dimensionality are limited to the linear correlation of the input data. Recently, a Convolutional Neural Network (CNN) Encoder was used to overcome this situation since it can use both linear and non-linear information to data reduction. This paper analyzes in-depth the performance to eliminate the useless paths using this CNN Encoder model. To measure the mentioned model efficiency, we combined it with different path planning algorithms. Next, the final algorithms (combined and not combined) are checked in a database that is composed of five scenarios. Each scenario contains fixed and dynamic obstacles. Their proposed model, the CNN Encoder, associated to other existent path planning algorithms in the literature, was able to obtain a time decrease to find the shortest path in comparison to all path planning algorithms analyzed. the average decreased time was 54.43 %.
△ Less
Submitted 5 August, 2020;
originally announced August 2020.
-
FCN+RL: A Fully Convolutional Network followed by Refinement Layers to Offline Handwritten Signature Segmentation
Authors:
Celso A. M. Lopes Junior,
Matheus Henrique M. da Silva,
Byron Leite Dantas Bezerra,
Bruno Jose Torres Fernandes,
Donato Impedovo
Abstract:
Although secular, handwritten signature is one of the most reliable biometric methods used by most countries. In the last ten years, the application of technology for verification of handwritten signatures has evolved strongly, including forensic aspects. Some factors, such as the complexity of the background and the small size of the region of interest - signature pixels - increase the difficulty…
▽ More
Although secular, handwritten signature is one of the most reliable biometric methods used by most countries. In the last ten years, the application of technology for verification of handwritten signatures has evolved strongly, including forensic aspects. Some factors, such as the complexity of the background and the small size of the region of interest - signature pixels - increase the difficulty of the targeting task. Other factors that make it challenging are the various variations present in handwritten signatures such as location, type of ink, color and type of pen, and the type of stroke. In this work, we propose an approach to locate and extract the pixels of handwritten signatures on identification documents, without any prior information on the location of the signatures. The technique used is based on a fully convolutional encoder-decoder network combined with a block of refinement layers for the alpha channel of the predicted image. The experimental results demonstrate that the technique outputs a clean signature with higher fidelity in the lines than the traditional approaches and preservation of the pertinent characteristics to the signer's spelling. To evaluate the quality of our proposal, we use the following image similarity metrics: SSIM, SIFT, and Dice Coefficient. The qualitative and quantitative results show a significant improvement in comparison with the baseline system.
△ Less
Submitted 28 May, 2020;
originally announced May 2020.
-
CNN Encoder to Reduce the Dimensionality of Data Image for Motion Planning
Authors:
Janderson Ferreira,
Agostinho A. F. Júnior,
Yves M. Galvão,
Bruno J. T. Fernandes,
Pablo Barros
Abstract:
Many real-world applications need path planning algorithms to solve tasks in different areas, such as social applications, autonomous cars, and tracking activities. And most importantly motion planning. Although the use of path planning is sufficient in most motion planning scenarios, they represent potential bottlenecks in large environments with dynamic changes. To tackle this problem, the numbe…
▽ More
Many real-world applications need path planning algorithms to solve tasks in different areas, such as social applications, autonomous cars, and tracking activities. And most importantly motion planning. Although the use of path planning is sufficient in most motion planning scenarios, they represent potential bottlenecks in large environments with dynamic changes. To tackle this problem, the number of possible routes could be reduced to make it easier for path planning algorithms to find the shortest path with less efforts. An traditional algorithm for path planning is the A*, it uses an heuristic to work faster than other solutions. In this work, we propose a CNN encoder capable of eliminating useless routes for motion planning problems, then we combine the proposed neural network output with A*. To measure the efficiency of our solution, we propose a database with different scenarios of motion planning problems. The evaluated metric is the number of the iterations to find the shortest path. The A* was compared with the CNN Encoder (proposal) with A*. In all evaluated scenarios, our solution reduced the number of iterations by more than 60\%.
△ Less
Submitted 10 April, 2020;
originally announced April 2020.
