Skip navigation
Please use this identifier to cite or link to this item: http://repository.iitr.ac.in/handle/123456789/16685
Full metadata record
DC FieldValueLanguage
dc.contributor.authorLubana E.S.-
dc.contributor.authorDIck R.P.-
dc.contributor.authorAggarwal V.-
dc.contributor.authorPradhan, Pyari Mohan-
dc.date.accessioned2020-12-02T14:17:24Z-
dc.date.available2020-12-02T14:17:24Z-
dc.date.issued2019-
dc.identifier.citationProceedings of International Conference on Image Processing, ICIP, (2019), 4165- 4169-
dc.identifier.isbn9.78154E+12-
dc.identifier.issn15224880-
dc.identifier.urihttps://doi.org/10.1109/ICIP.2019.8803645-
dc.identifier.urihttp://repository.iitr.ac.in/handle/123456789/16685-
dc.description.abstractIn-sensor energy-efficient deep learning accelerators have the potential to enable the use of deep neural networks in embedded vision applications. However, their negative impact on accuracy has been severely underestimated. The inference pipeline used in prior in-sensor deep learning accelerators bypasses the image signal processor (ISP), thereby disrupting the conventional vision pipeline and undermining accuracy of machine learning algorithms trained on conventional, post-ISP datasets. For example, the detection accuracy of an off-the-shelf Faster RCNN algorithm in a vehicle detection scenario reduces by 60%. To make in-sensor accelerators practical, we describe energy-efficient operations that yield most of the benefits of an ISP and reduce covariate shift between the training (ISP processed images) and target (RAW images) distributions. For the vehicle detection problem, our approach improves accuracy by 25-60%. Relative to the conventional ISP pipeline, energy consumption and response time improve by 30% and 34%, respectively. © 2019 IEEE.-
dc.description.sponsorshipThe Institute of Electrical and Electronics Engineers, Signal Processing Society-
dc.language.isoen_US-
dc.publisherIEEE Computer Society-
dc.relation.ispartofProceedings of International Conference on Image Processing, ICIP-
dc.subjectCovariate shift-
dc.subjectDeep learning accelerators-
dc.subjectImage signal processor-
dc.subjectRAW images-
dc.titleMinimalistic Image Signal Processing for Deep Learning Applications-
dc.typeConference Paper-
dc.scopusid57202045941-
dc.scopusid7202246612-
dc.scopusid56861825500-
dc.scopusid26639724100-
dc.affiliationLubana, E.S., University of Michigan, Ann Arbor, United States, Indian Institute of Technology, Roorkee, India-
dc.affiliationDIck, R.P., University of Michigan, Ann Arbor, United States-
dc.affiliationAggarwal, V., Indian Institute of Technology, Roorkee, India-
dc.affiliationPradhan, P.M., Indian Institute of Technology, Roorkee, India-
dc.identifier.conferencedetails26th IEEE International Conference on Image Processing, ICIP 2019, 22-25 September 2019-
Appears in Collections:Conference Publications [ECE]

Files in This Item:
There are no files associated with this item.
Show simple item record


Items in Repository are protected by copyright, with all rights reserved, unless otherwise indicated.