Convolutional Mean: A Simple Convolutional Neural Network for Illuminant Estimation

Han Gong

Convolutional Mean: A Simple Convolutional Neural Network for Illuminant Estimation

Han Gong

Research output: Chapter in Book/Report/Conference proceeding › Conference contribution

32 Downloads (Pure)

Abstract

We present Convolutional Mean (CM) – a simple and fast convolutional neural network for illuminant estimation. Our proposed method only requires a small neural network model (1.1K parameters) and a 48 × 32 thumbnail input image. Our unoptimized Python implementation takes 1 ms/image, which is arguably 3-3750× faster than the current leading solutions with similar accuracy. Using two public datasets, we show that our proposed light-weight method offers accuracy comparable to the current leading methods’ (which consist of thousands/millions of parameters) across several measures.

Original language	English
Title of host publication	British Machine Vision Conference (BMVC) 2019
Publisher	BMVA Press
Publication status	Published - Sep 2019
Event	British Machine Vision Conference - Duration: 9 Sep 2019 → 12 Sep 2019

Conference

Conference	British Machine Vision Conference
Period	9/09/19 → 12/09/19

Keywords

illuminant estimation
convolution

Access to Document

BMVC_2019Final published version, 8.08 MB

Cite this

@inproceedings{53d8022102824918a87a335e1ac00e69,

title = "Convolutional Mean: A Simple Convolutional Neural Network for Illuminant Estimation",

abstract = "We present Convolutional Mean (CM) – a simple and fast convolutional neural network for illuminant estimation. Our proposed method only requires a small neural network model (1.1K parameters) and a 48 × 32 thumbnail input image. Our unoptimized Python implementation takes 1 ms/image, which is arguably 3-3750× faster than the current leading solutions with similar accuracy. Using two public datasets, we show that our proposed light-weight method offers accuracy comparable to the current leading methods{\textquoteright} (which consist of thousands/millions of parameters) across several measures.",

keywords = "illuminant estimation, convolution",

author = "Han Gong",

year = "2019",

month = sep,

language = "English",

booktitle = "British Machine Vision Conference (BMVC) 2019",

publisher = "BMVA Press",

note = "British Machine Vision Conference ; Conference date: 09-09-2019 Through 12-09-2019",

}

TY - GEN

T1 - Convolutional Mean: A Simple Convolutional Neural Network for Illuminant Estimation

AU - Gong, Han

PY - 2019/9

Y1 - 2019/9

N2 - We present Convolutional Mean (CM) – a simple and fast convolutional neural network for illuminant estimation. Our proposed method only requires a small neural network model (1.1K parameters) and a 48 × 32 thumbnail input image. Our unoptimized Python implementation takes 1 ms/image, which is arguably 3-3750× faster than the current leading solutions with similar accuracy. Using two public datasets, we show that our proposed light-weight method offers accuracy comparable to the current leading methods’ (which consist of thousands/millions of parameters) across several measures.

AB - We present Convolutional Mean (CM) – a simple and fast convolutional neural network for illuminant estimation. Our proposed method only requires a small neural network model (1.1K parameters) and a 48 × 32 thumbnail input image. Our unoptimized Python implementation takes 1 ms/image, which is arguably 3-3750× faster than the current leading solutions with similar accuracy. Using two public datasets, we show that our proposed light-weight method offers accuracy comparable to the current leading methods’ (which consist of thousands/millions of parameters) across several measures.

KW - illuminant estimation

KW - convolution

M3 - Conference contribution

BT - British Machine Vision Conference (BMVC) 2019

PB - BMVA Press

T2 - British Machine Vision Conference

Y2 - 9 September 2019 through 12 September 2019

ER -