Deep fried convnets
… We also compare against a post-processed version of our model, where we train a deep
fried convnet and then apply SVD plus fine-tuning to the final softmax layer, which further re…
fried convnet and then apply SVD plus fine-tuning to the final softmax layer, which further re…
Deep simnets
… ConvNets we call Similarity Networks (SimNets), that preserves the simplicity and effectiveness
of ConvNets, … for designing deep networks with a higher abstraction level than ConvNets, …
of ConvNets, … for designing deep networks with a higher abstraction level than ConvNets, …
Deepström networks
… For the Fastfood approximation in Deep Fried Convnets we consider that φff is gained with
one stack of random features to form V in equation 3, except in the experiments of section 4.3 …
one stack of random features to form V in equation 3, except in the experiments of section 4.3 …
Fried binary embedding for high-dimensional visual features
… To address the two challenges above, we propose a novel approach, Fried Binary … We
call our approach as Fried Binary Embedding (FBE) following Deep Fried Convnets [24] …
call our approach as Fried Binary Embedding (FBE) following Deep Fried Convnets [24] …
Compressing convolutional neural networks
… We evaluate our compression scheme on eight deep learning image benchmark data sets
and compare against four competitive baselines. Although all compression schemes lead to …
and compare against four competitive baselines. Although all compression schemes lead to …
Fried binary embedding: From high-dimensional visual features to high-dimensional binary codes
… We call our approaches as Fried Binary Embedding following Deep Fried Convnets [17]
and Circulant Binary Embedding [12]. Extensive experiments show that our approach not only …
and Circulant Binary Embedding [12]. Extensive experiments show that our approach not only …
Compressing convolutional neural networks in the frequency domain
… We evaluate our compression scheme on eight deep learning image benchmark data sets
and compare against four competitive baselines. Although all compression schemes lead to …
and compare against four competitive baselines. Although all compression schemes lead to …
Deep networks with adaptive nyström approximation
… layers, ie these are classical convnets architectures ; (2) Deep Fried implements the Fastfood
… For the Fastfood approximation in Deep Fried Convnets we consider that φff is gained with …
… For the Fastfood approximation in Deep Fried Convnets we consider that φff is gained with …
Building efficient deep neural networks with unitary group convolutions
We propose unitary group convolutions (UGConvs), a building block for CNNs which
compose a group convolution with unitary transforms in feature space to learn a richer set of …
compose a group convolution with unitary transforms in feature space to learn a richer set of …
Food detection and recognition using convolutional neural network
In this paper, we apply a convolutional neural network (CNN) to the tasks of detecting and
recognizing food images. Because of the wide diversity of types of food, image recognition of …
recognizing food images. Because of the wide diversity of types of food, image recognition of …