All Projects → abhaydoke09 → Bilinear Cnn Tensorflow

abhaydoke09 / Bilinear Cnn Tensorflow

This is an implementation of Bilinear CNN for fine grained visual recognition using TensorFlow.

Programming Languages

python
139335 projects - #7 most used programming language

Projects that are alternatives of or similar to Bilinear Cnn Tensorflow

Tf Mobilenet V2
Mobilenet V2(Inverted Residual) Implementation & Trained Weights Using Tensorflow
Stars: ✭ 85 (-54.55%)
Mutual labels:  convolutional-neural-networks, cnn, deeplearning
Pytorch convlstm
convolutional lstm implementation in pytorch
Stars: ✭ 126 (-32.62%)
Mutual labels:  convolutional-neural-networks, cnn
Motionblur Detection By Cnn
Stars: ✭ 126 (-32.62%)
Mutual labels:  convolutional-neural-networks, cnn
Anime4k
A High-Quality Real Time Upscaler for Anime Video
Stars: ✭ 14,083 (+7431.02%)
Mutual labels:  convolutional-neural-networks, cnn
Deeplearning python
Deep Learning--深度学习
Stars: ✭ 123 (-34.22%)
Mutual labels:  cnn, deeplearning
Hyperdensenet
This repository contains the code of HyperDenseNet, a hyper-densely connected CNN to segment medical images in multi-modal image scenarios.
Stars: ✭ 124 (-33.69%)
Mutual labels:  convolutional-neural-networks, cnn
Image classifier
CNN image classifier implemented in Keras Notebook 🖼️.
Stars: ✭ 139 (-25.67%)
Mutual labels:  convolutional-neural-networks, cnn
All Conv Keras
All Convolutional Network: (https://arxiv.org/abs/1412.6806#) implementation in Keras
Stars: ✭ 115 (-38.5%)
Mutual labels:  convolutional-neural-networks, deeplearning
Simple cnn
Simple Convolutional Neural Network Library
Stars: ✭ 158 (-15.51%)
Mutual labels:  convolutional-neural-networks, cnn
Motion Sense
MotionSense Dataset for Human Activity and Attribute Recognition ( time-series data generated by smartphone's sensors: accelerometer and gyroscope)
Stars: ✭ 159 (-14.97%)
Mutual labels:  convolutional-neural-networks, deeplearning
Tf Adnet Tracking
Deep Object Tracking Implementation in Tensorflow for 'Action-Decision Networks for Visual Tracking with Deep Reinforcement Learning(CVPR 2017)'
Stars: ✭ 162 (-13.37%)
Mutual labels:  convolutional-neural-networks, cnn
Lenet 5
PyTorch implementation of LeNet-5 with live visualization
Stars: ✭ 122 (-34.76%)
Mutual labels:  convolutional-neural-networks, cnn
Keras transfer cifar10
Object classification with CIFAR-10 using transfer learning
Stars: ✭ 120 (-35.83%)
Mutual labels:  convolutional-neural-networks, cnn
Deeplearning Notes
Notes for Deep Learning Specialization Courses led by Andrew Ng.
Stars: ✭ 126 (-32.62%)
Mutual labels:  cnn, deeplearning
Deepway
This project is an aid to the blind. Till date there has been no technological advancement in the way the blind navigate. So I have used deep learning particularly convolutional neural networks so that they can navigate through the streets.
Stars: ✭ 118 (-36.9%)
Mutual labels:  convolutional-neural-networks, cnn
Visualizingcnn
🙈A PyTorch implementation of the paper "Visualizing and Understanding Convolutional Networks." (ECCV 2014)
Stars: ✭ 132 (-29.41%)
Mutual labels:  convolutional-neural-networks, deeplearning
Cnnimageretrieval
CNN Image Retrieval in MatConvNet: Training and evaluating CNNs for Image Retrieval in MatConvNet
Stars: ✭ 168 (-10.16%)
Mutual labels:  convolutional-neural-networks, cnn
Har Keras Cnn
Human Activity Recognition (HAR) with 1D Convolutional Neural Network in Python and Keras
Stars: ✭ 97 (-48.13%)
Mutual labels:  convolutional-neural-networks, deeplearning
Deepgaze
Computer Vision library for human-computer interaction. It implements Head Pose and Gaze Direction Estimation Using Convolutional Neural Networks, Skin Detection through Backprojection, Motion Detection and Tracking, Saliency Map.
Stars: ✭ 1,552 (+729.95%)
Mutual labels:  convolutional-neural-networks, cnn
Livianet
This repository contains the code of LiviaNET, a 3D fully convolutional neural network that was employed in our work: "3D fully convolutional networks for subcortical segmentation in MRI: A large-scale study"
Stars: ✭ 143 (-23.53%)
Mutual labels:  convolutional-neural-networks, cnn

Bilinear_CNN TensorFlow

This is an implementation of Bilinear Convolutional Neural Network (Bilinear_CNN) using TensorFlow.

Main objective of this project is to implement Bilinear Convolutional Neural Network (Bilinear_CNN) for Fine-grained Visual Recognition using TensorFlow. I implemented the Bilinear Convolutional Neural Network (Bilinear_CNN) model as described in the http://vis-www.cs.umass.edu/bcnn/docs/bcnn_iccv15.pdf and trained it on the FGVC-Aircraft dataset with 100 categories. Bilinear Convolutional Neural Network model combines two Convolutional Neural Network architectures pre-trained on the ImageNet dataset using outer product at each location in the image. Training Bilinear Convolutional Neural Network model is a two step training procedure in which the last fully connected layer is trained first followed by the fine-tuning entire model using back propagation. In this project, I present experimental results of two methods on top of the Bilinear_CNN (DD) model as described in http://vis-www.cs.umass.edu/bcnn/docs/bcnn_iccv15.pdf which uses two VGG16 models pretrained on ImageNet dataset and then Bilinear_CNN (DD) model is trained on the FGVCAircraft dataset. One, I experimented with a slightly different approach in the two-step training procedure, where the training of the last layer is interrupted after 10-15 epochs and fine-tuning the entire model was started after that. Two, I used random cropping of images during the training of Bilinear_CNN (DD) model to see if there is any significant improvement in the accuracy of the Bilinear_CNN (DD) model on the FGVC-Aircraft dataset. I obtain 86.4% accuracy with the first method and 85.41% accuracy with the second method. Training of the network is done on NVIDIA Tesla M40 GPU. Training of last layer of Bilinear_CNN (DD) TensorFlow model runs at 20 frames/second and fine-tuning the entire Bilinear_CNN (DD) TensorFlow model runs at 10 frames/second.

To download the VGG16 model weigths and to get the TensorFlow model for VGG16, go to -> https://www.cs.toronto.edu/~frossard/post/vgg16/

I provide the Bilinear_CNN (BCNN) implmentation in TensorFlow.

bcnn_DD_woft.py and bcnn_DD_woft_with_random_crops.py are TensorFlow files used for the first step of the training procedure where only last layer of the Bilinear_CNN (DD) model is trained. --> Learning rate = 0.9 --> Optimizer = Momentum optimizer with 0.9 momentum

bcnn_finetuning.py and bcnn_finetuning_with_random_crops.py are TensorFlow files used for the second step of the training procedure where finetuning is performed on the entire Bilinear_CNN (DD) model. --> Learning rate = 0.001 --> Optimizer = Momentum optimizer with 0.9 momentum

Note that the project description data, including the texts, logos, images, and/or trademarks, for each open source project belongs to its rightful owner. If you wish to add or remove any projects, please contact us at [email protected].