Exploring task-agnostic, ShapeNet-based object recognition for mobile robots

Agnese Chiatti, Gianluca Bardaro, Emanuele Bastianelli, Ilaria Tiddi, Prasenjit Mitra, Enrico Motta

Research output: Chapter in Book / Report / Conference proceedingChapterAcademicpeer-review

Abstract

This position paper presents an attempt to improve the scalability of existing object recognition methods, which largely rely on supervision and imply a huge availability of manually-labelled data points. Moreover, in the context of mobile robotics, data sets and experimental settings are often handcrafted based on the specific task the object recognition is aimed at, e.g. object grasping. In this work, we argue instead that publicly available open data such as ShapeNet [8] can be used for object classification first, and then to link objects to their related concepts, leading to task-agnostic knowledge acquisition practices. To this aim, we evaluated five pipelines for object recognition, where target classes were all entities collected from ShapeNet and matching was based on: (i) shape-only features, (ii) RGB histogram comparison, (iii) a combination of shape and colour matching, (iv) image feature descriptors, and (v) inexact, normalised cross-correlation, resembling the Deep, Siamese-like NN architecture of [31]. We discussed the relative impact of shape-derived and colour-derived features, as well as suitability of all tested solutions for future application to real-life use cases.

Original languageEnglish
Title of host publicationProceedings of the Workshops of the EDBT/ICDT 2019 Joint Conference (EDBT/ICDT 2019) Lisbon, Portugal, March 26, 2019
EditorsPaolo Papotti
PublisherCEUR Workshop Proceedings
Pages1-8
Number of pages8
Publication statusPublished - 2019
Event2019 Workshops of the EDBT/ICDT Joint Conference, EDBT/ICDT-WS 2019 - Lisbon, Portugal
Duration: 26 Mar 2019 → …

Publication series

NameCEUR Workshop Proceedings
PublisherCEUR Workshop Proceedings
Volume2322
ISSN (Print)1613-0073

Conference

Conference2019 Workshops of the EDBT/ICDT Joint Conference, EDBT/ICDT-WS 2019
CountryPortugal
CityLisbon
Period26/03/19 → …

Fingerprint

Object recognition
Mobile robots
Color matching
Knowledge acquisition
Scalability
Robotics
Pipelines
Availability
Color

Bibliographical note

DARLI-AP: Data Analytics Solutions for Real-Life Applications

Cite this

Chiatti, A., Bardaro, G., Bastianelli, E., Tiddi, I., Mitra, P., & Motta, E. (2019). Exploring task-agnostic, ShapeNet-based object recognition for mobile robots. In P. Papotti (Ed.), Proceedings of the Workshops of the EDBT/ICDT 2019 Joint Conference (EDBT/ICDT 2019) Lisbon, Portugal, March 26, 2019 (pp. 1-8). (CEUR Workshop Proceedings; Vol. 2322). CEUR Workshop Proceedings.
Chiatti, Agnese ; Bardaro, Gianluca ; Bastianelli, Emanuele ; Tiddi, Ilaria ; Mitra, Prasenjit ; Motta, Enrico. / Exploring task-agnostic, ShapeNet-based object recognition for mobile robots. Proceedings of the Workshops of the EDBT/ICDT 2019 Joint Conference (EDBT/ICDT 2019) Lisbon, Portugal, March 26, 2019. editor / Paolo Papotti. CEUR Workshop Proceedings, 2019. pp. 1-8 (CEUR Workshop Proceedings).
@inbook{b48d6538e6b84d75afeb3d77677fd232,
title = "Exploring task-agnostic, ShapeNet-based object recognition for mobile robots",
abstract = "This position paper presents an attempt to improve the scalability of existing object recognition methods, which largely rely on supervision and imply a huge availability of manually-labelled data points. Moreover, in the context of mobile robotics, data sets and experimental settings are often handcrafted based on the specific task the object recognition is aimed at, e.g. object grasping. In this work, we argue instead that publicly available open data such as ShapeNet [8] can be used for object classification first, and then to link objects to their related concepts, leading to task-agnostic knowledge acquisition practices. To this aim, we evaluated five pipelines for object recognition, where target classes were all entities collected from ShapeNet and matching was based on: (i) shape-only features, (ii) RGB histogram comparison, (iii) a combination of shape and colour matching, (iv) image feature descriptors, and (v) inexact, normalised cross-correlation, resembling the Deep, Siamese-like NN architecture of [31]. We discussed the relative impact of shape-derived and colour-derived features, as well as suitability of all tested solutions for future application to real-life use cases.",
author = "Agnese Chiatti and Gianluca Bardaro and Emanuele Bastianelli and Ilaria Tiddi and Prasenjit Mitra and Enrico Motta",
note = "DARLI-AP: Data Analytics Solutions for Real-Life Applications",
year = "2019",
language = "English",
series = "CEUR Workshop Proceedings",
publisher = "CEUR Workshop Proceedings",
pages = "1--8",
editor = "Paolo Papotti",
booktitle = "Proceedings of the Workshops of the EDBT/ICDT 2019 Joint Conference (EDBT/ICDT 2019) Lisbon, Portugal, March 26, 2019",

}

Chiatti, A, Bardaro, G, Bastianelli, E, Tiddi, I, Mitra, P & Motta, E 2019, Exploring task-agnostic, ShapeNet-based object recognition for mobile robots. in P Papotti (ed.), Proceedings of the Workshops of the EDBT/ICDT 2019 Joint Conference (EDBT/ICDT 2019) Lisbon, Portugal, March 26, 2019. CEUR Workshop Proceedings, vol. 2322, CEUR Workshop Proceedings, pp. 1-8, 2019 Workshops of the EDBT/ICDT Joint Conference, EDBT/ICDT-WS 2019, Lisbon, Portugal, 26/03/19.

Exploring task-agnostic, ShapeNet-based object recognition for mobile robots. / Chiatti, Agnese; Bardaro, Gianluca; Bastianelli, Emanuele; Tiddi, Ilaria; Mitra, Prasenjit; Motta, Enrico.

Proceedings of the Workshops of the EDBT/ICDT 2019 Joint Conference (EDBT/ICDT 2019) Lisbon, Portugal, March 26, 2019. ed. / Paolo Papotti. CEUR Workshop Proceedings, 2019. p. 1-8 (CEUR Workshop Proceedings; Vol. 2322).

Research output: Chapter in Book / Report / Conference proceedingChapterAcademicpeer-review

TY - CHAP

T1 - Exploring task-agnostic, ShapeNet-based object recognition for mobile robots

AU - Chiatti, Agnese

AU - Bardaro, Gianluca

AU - Bastianelli, Emanuele

AU - Tiddi, Ilaria

AU - Mitra, Prasenjit

AU - Motta, Enrico

N1 - DARLI-AP: Data Analytics Solutions for Real-Life Applications

PY - 2019

Y1 - 2019

N2 - This position paper presents an attempt to improve the scalability of existing object recognition methods, which largely rely on supervision and imply a huge availability of manually-labelled data points. Moreover, in the context of mobile robotics, data sets and experimental settings are often handcrafted based on the specific task the object recognition is aimed at, e.g. object grasping. In this work, we argue instead that publicly available open data such as ShapeNet [8] can be used for object classification first, and then to link objects to their related concepts, leading to task-agnostic knowledge acquisition practices. To this aim, we evaluated five pipelines for object recognition, where target classes were all entities collected from ShapeNet and matching was based on: (i) shape-only features, (ii) RGB histogram comparison, (iii) a combination of shape and colour matching, (iv) image feature descriptors, and (v) inexact, normalised cross-correlation, resembling the Deep, Siamese-like NN architecture of [31]. We discussed the relative impact of shape-derived and colour-derived features, as well as suitability of all tested solutions for future application to real-life use cases.

AB - This position paper presents an attempt to improve the scalability of existing object recognition methods, which largely rely on supervision and imply a huge availability of manually-labelled data points. Moreover, in the context of mobile robotics, data sets and experimental settings are often handcrafted based on the specific task the object recognition is aimed at, e.g. object grasping. In this work, we argue instead that publicly available open data such as ShapeNet [8] can be used for object classification first, and then to link objects to their related concepts, leading to task-agnostic knowledge acquisition practices. To this aim, we evaluated five pipelines for object recognition, where target classes were all entities collected from ShapeNet and matching was based on: (i) shape-only features, (ii) RGB histogram comparison, (iii) a combination of shape and colour matching, (iv) image feature descriptors, and (v) inexact, normalised cross-correlation, resembling the Deep, Siamese-like NN architecture of [31]. We discussed the relative impact of shape-derived and colour-derived features, as well as suitability of all tested solutions for future application to real-life use cases.

UR - http://www.scopus.com/inward/record.url?scp=85062646474&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=85062646474&partnerID=8YFLogxK

M3 - Chapter

T3 - CEUR Workshop Proceedings

SP - 1

EP - 8

BT - Proceedings of the Workshops of the EDBT/ICDT 2019 Joint Conference (EDBT/ICDT 2019) Lisbon, Portugal, March 26, 2019

A2 - Papotti, Paolo

PB - CEUR Workshop Proceedings

ER -

Chiatti A, Bardaro G, Bastianelli E, Tiddi I, Mitra P, Motta E. Exploring task-agnostic, ShapeNet-based object recognition for mobile robots. In Papotti P, editor, Proceedings of the Workshops of the EDBT/ICDT 2019 Joint Conference (EDBT/ICDT 2019) Lisbon, Portugal, March 26, 2019. CEUR Workshop Proceedings. 2019. p. 1-8. (CEUR Workshop Proceedings).