Age Estimation Using Trainable Gabor Wavelet Layers In A Convolutional Neural Network

kwon, hyuk jin; Koo, Hyung Il; Soh, Jae Woong; Cho, Nam Ik

doi:10.1109/icip.2019.8803442

Cited by 4 publications

(18 citation statements)

References 22 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…Using Gabor wavelets with CNNs have received a tremendous attention as well [9,[41][42][43]. [41] use a Gabor filter bank as the first layer of a CNN and the bank gets updated using the standard back-propagation network leaning phase.…”

Section: Related Workmentioning

confidence: 99%

“…Authors in [43] introduce a Gabor Neural Network (GNN) where Gabor filters are incorporated into the convolution filter as a modulation process, in a spirit similar to the above mentioned works. In contrast to the above works where fixed Gabor filters are used [9], introduce a trainable Gabor wavelets (TGW) layer. The authors present a method where the hyperparameters of the wavelets are learned from the input and a novel 1 × 1 convolution layers are employed to create steerable filters.…”

Section: Related Workmentioning

confidence: 99%

“…. ., bzc, and by just varying bzc, variety of sampling grids can be achieved [9]. For a loss function L, we need to compute @L @z in order to train for the wavelets layer that is cascaded with our CNN.…”

Section: Gabor Wavelets Layermentioning

confidence: 99%

“…To use the steering property, where a linear combination of finite set of responses can be used to represent convolution at any orientation, a 1 × 1 convolution layer, working as a linear combination layer, is added to the output of the generated filters. For this layer, ten equally spaced fixed orientations are selected, working as basis filters: 9˚, 27˚, 45˚, 63˚, 81˚, 99˚, 117˚, 135˚, 153˚, and 171˚ [9].…”

Section: Gabor Wavelets Layermentioning

confidence: 99%

“…However, we adopt an approach where the convolutional network is employed to learn the wavelet parameters along with learning the dataset. These Trainable Gabor wavelets (TGW) [9] make up for the backbone of our network. Each TGW accepts a single channel input, with a multi-channel output, and learns the best parameters to generate a set of Gabor filters.…”

Section: Introductionmentioning

confidence: 99%

See 4 more Smart Citations

Pedestrian attribute recognition using two-branch trainable Gabor wavelets network

Junejo

2021

PLoS ONE

View full text Add to dashboard Cite

Keeping an eye on pedestrians as they navigate through a scene, surveillance cameras are everywhere. With this context, our paper addresses the problem of pedestrian attribute recognition (PAR). This problem entails recognizing attributes such as age-group, clothing style, accessories, footwear style etc. This multi-label problem is extremely challenging even for human observers and has rightly garnered attention from the computer vision community. Towards a solution to this problem, in this paper, we adopt trainable Gabor wavelets (TGW) layers and cascade them with a convolution neural network (CNN). Whereas other researchers are using fixed Gabor filters with the CNN, the proposed layers are learnable and adapt to the dataset for a better recognition. We propose a two-branch neural network where mixed layers, a combination of the TGW and convolutional layers, make up the building block of our deep neural network. We test our method on twoo challenging publicly available datasets and compare our results with state of the art.

show abstract

Section: Related Workmentioning

confidence: 99%

Section: Related Workmentioning

confidence: 99%

Section: Gabor Wavelets Layermentioning

confidence: 99%

Section: Gabor Wavelets Layermentioning

confidence: 99%

Section: Introductionmentioning

confidence: 99%

See 3 more Smart Citations

Pedestrian attribute recognition using two-branch trainable Gabor wavelets network

Junejo

2021

PLoS ONE

View full text Add to dashboard Cite

show abstract

Pedestrian attribute recognition using trainable Gabor wavelets

Junejo

Ahmed

Lataifeh

2021

Heliyon

View full text Add to dashboard Cite

Surveillance cameras are everywhere keeping an eye on pedestrians or people as they navigate through the scene. Within this context, our paper addresses the problem of pedestrian attribute recognition (PAR). This problem entails the extraction of different attributes such as age-group, clothing style, accessories, footwear style etc. This is a multi-label problem with a host of challenges even for human observers. As such, the topic has rightly attracted attention recently. In this work, we integrate trainable Gabor wavelet (TGW) layers inside a convolution neural network (CNN). Whereas other researchers have used fixed Gabor filters with the CNN, the proposed layers are learnable and adapt to the dataset for a better recognition. We test our method on publicly available challenging datasets and demonstrate considerable improvements over state of the art approaches.

show abstract

Multi-Branch Gabor Wavelet Layers for Pedestrian Attribute Recognition

Junejo

2021

IEEE Access

View full text Add to dashboard Cite

Age Estimation Using Trainable Gabor Wavelet Layers In A Convolutional Neural Network

Cited by 4 publications

References 22 publications

Pedestrian attribute recognition using two-branch trainable Gabor wavelets network

Pedestrian attribute recognition using two-branch trainable Gabor wavelets network

Pedestrian attribute recognition using trainable Gabor wavelets

Multi-Branch Gabor Wavelet Layers for Pedestrian Attribute Recognition

Contact Info

Product

Resources

About