Abstract
This paper introduces a deep learning enabled generative sensing framework which integrates low-end sensors with computational intelligence to attain a high recognition accuracy on par with that attained with high-end sensors. The proposed generative sensing framework aims at transforming low-end, low-quality sensor data into higher quality sensor data in terms of achieved classification accuracy. The low-end data can be transformed into higher quality data of the same modality or into data of another modality. Different from existing methods for image generation, the proposed framework is based on discriminative models and targets to maximize the recognition accuracy rather than a similarity measure. This is achieved through the introduction of selective feature regeneration in a deep neural network (DNN). The proposed generative sensing will essentially transform low-quality sensor data into high-quality information for robust perception. Results are presented to illustrate the performance of the proposed framework.
| Original language | English (US) |
|---|---|
| Title of host publication | Proceedings - IEEE 1st Conference on Multimedia Information Processing and Retrieval, MIPR 2018 |
| Publisher | Institute of Electrical and Electronics Engineers Inc. |
| Pages | 100-105 |
| Number of pages | 6 |
| ISBN (Electronic) | 9781538618578 |
| DOIs | |
| State | Published - Jun 26 2018 |
| Event | 1st IEEE Conference on Multimedia Information Processing and Retrieval, MIPR 2018 - Miami, United States Duration: Apr 10 2018 → Apr 12 2018 |
Other
| Other | 1st IEEE Conference on Multimedia Information Processing and Retrieval, MIPR 2018 |
|---|---|
| Country | United States |
| City | Miami |
| Period | 4/10/18 → 4/12/18 |
Fingerprint
Keywords
- classification
- deep learning
- generative sensing
- image
- infrared
- IR
- multi-modal
- NIR
- quality
- recognition
- resolution
- visible
ASJC Scopus subject areas
- Computer Networks and Communications
- Signal Processing
- Media Technology
Cite this
Generative Sensing : Transforming Unreliable Sensor Data for Reliable Recognition. / Karam, Lina; Borkar, Tejas; Cao, Yu; Chae, Junseok.
Proceedings - IEEE 1st Conference on Multimedia Information Processing and Retrieval, MIPR 2018. Institute of Electrical and Electronics Engineers Inc., 2018. p. 100-105.Research output: Chapter in Book/Report/Conference proceeding › Conference contribution
}
TY - GEN
T1 - Generative Sensing
T2 - Transforming Unreliable Sensor Data for Reliable Recognition
AU - Karam, Lina
AU - Borkar, Tejas
AU - Cao, Yu
AU - Chae, Junseok
PY - 2018/6/26
Y1 - 2018/6/26
N2 - This paper introduces a deep learning enabled generative sensing framework which integrates low-end sensors with computational intelligence to attain a high recognition accuracy on par with that attained with high-end sensors. The proposed generative sensing framework aims at transforming low-end, low-quality sensor data into higher quality sensor data in terms of achieved classification accuracy. The low-end data can be transformed into higher quality data of the same modality or into data of another modality. Different from existing methods for image generation, the proposed framework is based on discriminative models and targets to maximize the recognition accuracy rather than a similarity measure. This is achieved through the introduction of selective feature regeneration in a deep neural network (DNN). The proposed generative sensing will essentially transform low-quality sensor data into high-quality information for robust perception. Results are presented to illustrate the performance of the proposed framework.
AB - This paper introduces a deep learning enabled generative sensing framework which integrates low-end sensors with computational intelligence to attain a high recognition accuracy on par with that attained with high-end sensors. The proposed generative sensing framework aims at transforming low-end, low-quality sensor data into higher quality sensor data in terms of achieved classification accuracy. The low-end data can be transformed into higher quality data of the same modality or into data of another modality. Different from existing methods for image generation, the proposed framework is based on discriminative models and targets to maximize the recognition accuracy rather than a similarity measure. This is achieved through the introduction of selective feature regeneration in a deep neural network (DNN). The proposed generative sensing will essentially transform low-quality sensor data into high-quality information for robust perception. Results are presented to illustrate the performance of the proposed framework.
KW - classification
KW - deep learning
KW - generative sensing
KW - image
KW - infrared
KW - IR
KW - multi-modal
KW - NIR
KW - quality
KW - recognition
KW - resolution
KW - visible
UR - http://www.scopus.com/inward/record.url?scp=85050122354&partnerID=8YFLogxK
UR - http://www.scopus.com/inward/citedby.url?scp=85050122354&partnerID=8YFLogxK
U2 - 10.1109/MIPR.2018.00025
DO - 10.1109/MIPR.2018.00025
M3 - Conference contribution
AN - SCOPUS:85050122354
SP - 100
EP - 105
BT - Proceedings - IEEE 1st Conference on Multimedia Information Processing and Retrieval, MIPR 2018
PB - Institute of Electrical and Electronics Engineers Inc.
ER -