Human and object recognition with a high-resolution tactile sensor

Gandarias, Juan M.; Gómez-de-Gabriel, Jesús M.; García-Cerezo, Alfonso J.

doi:10.1109/icsens.2017.8234203

Cited by 35 publications

(33 citation statements)

References 15 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…To have a comparison point to our proposed method (DCNN), the DCNN-SVM method was also implemented. This method was previously presented in [ 53 ], where the first pre-trained layers of a DCNN are used to extract the features of the input tactile images. Then, the SVM replaces the last layer of the network and must be trained with pre-labelled tactile images.…”

Section: Methodsmentioning

confidence: 99%

“…The recognition procedure adopted in this paper complies with the idea of classifying an object by its contact shape, treating pressure data from a high-resolution tactile sensor as common images, and using a deep convolutional neural network [ 53 ]. This method was successfully implemented with a tactile sensor mounted on a rigid probe, applied to robotics search and rescue tasks [ 54 ], showing how tactile perception provides valuable information for the search of potential victims—especially in low-visibility scenarios.…”

Section: Tactile Recognitionmentioning

confidence: 99%

See 1 more Smart Citation

Enhancing Perception with Tactile Object Recognition in Adaptive Grippers for Human–Robot Interaction

Gandarias

Gómez-de-Gabriel

García-Cerezo

2018

Sensors

View full text Add to dashboard Cite

The use of tactile perception can help first response robotic teams in disaster scenarios, where visibility conditions are often reduced due to the presence of dust, mud, or smoke, distinguishing human limbs from other objects with similar shapes. Here, the integration of the tactile sensor in adaptive grippers is evaluated, measuring the performance of an object recognition task based on deep convolutional neural networks (DCNNs) using a flexible sensor mounted in adaptive grippers. A total of 15 classes with 50 tactile images each were trained, including human body parts and common environment objects, in semi-rigid and flexible adaptive grippers based on the fin ray effect. The classifier was compared against the rigid configuration and a support vector machine classifier (SVM). Finally, a two-level output network has been proposed to provide both object-type recognition and human/non-human classification. Sensors in adaptive grippers have a higher number of non-null tactels (up to 37% more), with a lower mean of pressure values (up to 72% less) than when using a rigid sensor, with a softer grip, which is needed in physical human–robot interaction (pHRI). A semi-rigid implementation with 95.13% object recognition rate was chosen, even though the human/non-human classification had better results (98.78%) with a rigid sensor.

show abstract

Section: Methodsmentioning

confidence: 99%

Section: Tactile Recognitionmentioning

confidence: 99%

Enhancing Perception with Tactile Object Recognition in Adaptive Grippers for Human–Robot Interaction

Gandarias

Gómez-de-Gabriel

García-Cerezo

2018

Sensors

View full text Add to dashboard Cite

show abstract

“…While most of the work done was focused on the methodology itself, few works addressed the implementation on embedded platforms where the real application should reside. Gandarias et al [11] used two approaches to classify eight objects: finger, hand, arm, pen, scissors, pliers, sticky tape, and Allen key, using a 28 × 50 tactile sensory array attached to a robotic arm, the first approach using the Speeded-Up Robust Features (SURF) descriptor, while the second a pre-trained AlexNet CNN for feature extraction, with a Support Vector Machine (SVM) classifier for both approaches. In Yuan et al's research [12], a CNN was also used for active tactile clothing perception, to classify clothes grasped by a robotic arm equipped with a tactile sensor that output a large RGB pressure map.…”

Section: State-of-the-artmentioning

confidence: 99%

Smart Tactile Sensing Systems Based on Embedded CNN Implementations

et al. 2020

View full text Add to dashboard Cite

Embedding machine learning methods into the data decoding units may enable the extraction of complex information making the tactile sensing systems intelligent. This paper presents and compares the implementations of a convolutional neural network model for tactile data decoding on various hardware platforms. Experimental results show comparable classification accuracy of 90.88% for Model 3, overcoming similar state-of-the-art solutions in terms of time inference. The proposed implementation achieves a time inference of 1.2 ms while consuming around 900 μ J. Such an embedded implementation of intelligent tactile data decoding algorithms enables tactile sensing systems in different application domains such as robotics and prosthetic devices.

show abstract

“…The proposed framework is based on deep Convolutional Neural Networks (CNNs). This choice was motivated by the good performance reported recently in multi-class tactile recognition [7], [8] when using CNNs. However, note that the use of CNNs for Zero-Shot Learning is not straightforward: if we simply train a CNN to map tactile data into object classes, the CNN will miss output classes having no training data.…”

Section: Introductionmentioning

confidence: 99%

A Deep Learning Framework for Tactile Recognition of Known as Well as Novel Objects

Abderrahmane

Ganesh

Crosnier

et al. 2020

IEEE Trans. Ind. Inf.

View full text Add to dashboard Cite

This paper addresses the recognition of daily-life objects by a robot equipped with tactile sensors. The main contribution is a deep learning framework that can recognize objects already touched as well as objects never touched before. To this end, we train a Deconvolutional Neural Network that generates synthetic tactile data for novel classes. Then, we use both these synthetic data and the real data collected by touching objects, to train a Convolutional Neural Network to recognize both known (trained) objects and novel objects. Furthermore, we propose a method for integrating newly encountered data into novel classes. Finally, we evaluate the framework using the largest available dataset of tactile objects descriptions.

show abstract

Human and object recognition with a high-resolution tactile sensor

Cited by 35 publications

References 15 publications

Enhancing Perception with Tactile Object Recognition in Adaptive Grippers for Human–Robot Interaction

Enhancing Perception with Tactile Object Recognition in Adaptive Grippers for Human–Robot Interaction

Smart Tactile Sensing Systems Based on Embedded CNN Implementations

A Deep Learning Framework for Tactile Recognition of Known as Well as Novel Objects

Contact Info

Product

Resources

About