Before we get into the CNN code, I would like to spend time in. The score of each class for an image is then obtained by comparing each template with the image using an inner product (or dot product) one by one to find the one that “fits” best. If there is a single class, the term "recognition" is often applied, whereas a multi-class recognition task is often called "classification". org/pdf/1702. However, originally, CNNs were designed to cope with image data. GitHub has ≈6M of active users (and 3M after some filtering). Image Recogniser (Low-level API Tutorial) SqueezeNet 1000 Classes Image Classifier This chapter introduces how to build an image classifier by using low-level API only. Now it is time to run the CNN (this can take a while…). Training set includes about 39000 images while test set has around 12000 images. and has an example on github here Some of his examples don't use tensor flow but the one I linked to github uses CNN for text classification using tensor flow share | improve this answer. 0 showing alternately the input image, an overlay of FCN-Alexnet predictions, an overlay of FCN-8s predictions and the ground truth. January 22, 2017. edu, [email protected]
Another way of using pre-trained CNNs for transfer learning is to fine-tune CNNs by initializing network weights from a pre-trained network and then re-training the network with the new dataset. R-CNN's have proved highly effective in detecting and classifying objects in natural images, achieving mAP scores far higher than previous techniques. In this vignette, we will describe the core functionality of the CNTK framework, and how to use it's R bindings to ingest data, train a model, and evaluate it on a test set. Personalized Image Classi cation from EEG Signals using Deep Learning A Degree Thesis Submitted to the Faculty of the Escola T ecnica d’Enginyeria de Telecomunicaci o de Barcelona. It takes a source image as input and predicts the likelihood of whether the target image is a real or fake image. 2 million images and 1000 classes) as a starting point. EXERCISE: Download the Street View House Numbers (SVHN) data set. reshape(x, [-1, 28, 28, 1]) # 첫번째 Convolution Layer # 5x5 Kernel Size를 가진 32개의 Filter를. This makes it extremely fast, more than 1000x faster than R-CNN and 100x faster than Fast R-CNN. Mask R-CNN Installation. In our blog post we will use the pretrained model to classify, annotate and segment images into these 1000 classes. One concern about LULC classification is that at high resolutions, there is a significant amount of variability in the data. png) ![Inria](images/inria-log. A deep CNN of Dan Ciresan et al. A few sample labeled images from the training dataset are shown below. This article will explain the Convolutional Neural Network (CNN) with an illustration of image classification. In fact, it is more natural to think of images as belonging to multiple classes rather than a single class. The image illustrates the task; To see if that is possible I downloaded the raw data from an FU-Berlin ftp- server. affiliations[ ![Heuritech](images/logo heuritech v2. The set of classes is very diverse. it can be used. The reason for not starting with an NLP task is that applying CNNs to NLP tasks (for example, sentence classification) is not very straightforward. It is guided by the contextual discriminator to produce real looking images. Visual attention has been successfully applied in structural prediction tasks such as visual captioning and question answering. And I've tested tensorflow verions 1. def CNN (input_dim, num_classes):. Image Classification Using Convolutional Neural Networks in TensorFlow; This blog post introduces a type of neural network called a convolutional neural network (CNN) using Python and TensorFlow. In this blog post, we will quickly understand how to use state-of-the-art Deep Learning models in Keras to solve a supervised image classification problem using our own dataset with/without GPU acceleration. Now that we have our images downloaded and organized, the next step is to train a Convolutional Neural Network (CNN) on top of the data. R-CNN (Girshick et al. on Computer Vision and Pattern Recognition (CVPR), Boston, 2015. As a result, we used CNN as the baseline for the experiments and tried. Given the role of apparel in society, CBIR of fashion images has many applications. However, most of these models only input the whole spectral bands into RNNs directly, which may not fully explore the specific properties of HSIs. handong1587's blog. Also called Softmax Loss. Furthermore, our architecture can process images of any resolution, unlike most existing approaches based on CNN. In this first post, I will look into how to use convolutional neural network to build a classifier, particularly Convolutional Neural Networks for Sentence Classification - Yoo Kim. Census income classification with scikit-learn - Using the standard adult census income dataset, this notebook trains a k-nearest neighbors classifier using scikit-learn and then explains predictions using shap. We first apply a number of convolutional layers to extract features from our image, and then we apply deconvolutional layers to upscale (increase the spacial resolution) of our features. Examples to use pre-trained CNNs for image classification and feature extraction. My approach here is a convolution Neural network to generate the color channels. This approach is compared to. In this article I'll explain the DNN approach, using the Keras code library. ] (Used in this work)(Enable a controlled comparison with prior detection work). head (obtained using a laser scanner) and image inten-sities were used for classifying gender. The score of each class for an image is then obtained by comparing each template with the image using an inner product (or dot product) one by one to find the one that “fits” best. Fullerton, CA, U. Topic-Enhanced Memory Networks for Personalised Point-of-Interest Recommendation arXiv_CV arXiv_CV Attention Relation Memory_Networks Recommendation. The aim of this post is to build a custom Mask R-CNN model that can detect the area of damage on a car (see the image example above). The well-known application of CNN is image classification, where a fixed dimension image is fed into a network along with different channels (RGB in the case of a color image) and after various steps of convolution, pooling and fully connected layers, network outputs class probabilities for the image. Below are a few examples of inferred alignments. In this tutorial, we will discuss how to use those models as a Feature Extractor and train a new model for a. reshape(x, [-1, 28, 28, 1]) # 첫번째 Convolution Layer # 5x5 Kernel Size를 가진 32개의 Filter를. Convolutioning an image with Gabor filters generates transformed images. Fork the CIFAR-10 tutorial and swap in the SVHN as the input data. We will classify images using a CNN. Classification + Localization. These 60,000 images are partitioned into a training. CNN for multi-class image recognition in tensorflow. Our focus in this year’s submissions is to provide the best-performing. com Image Classification Using Ensemble Deep CNN. of image set clustering for object classi cation as well as ne-grained classi cation. But I got more images of Where's Waldo from it. CNN 3D Images using Tensorflow. What the script does:. We want to use a sigmoid activation function on the neuron output. To see this, consider the extreme case where all 1. Image classification using CNN features and linear SVM - feature_vector_from_cnn. Loading Data into Keras Model. Before we get into the CNN code, I would like to spend time in. import tensorflow as tf # change this as you see fit image_path = sys. The difference here is that instead of using image features such as HOG or SURF, features are extracted using a CNN. The figure below provides the CNN model architecture that we are going to implement using Tensorflow. I use ADAM as the optimizer, select binary cross-entropy as the loss function for this binary classification task and use accuracy to evaluate the results (there is no need to use a different metric because in this particular case accuracy is what I am interested in). In this case, we will use a Mask R-CNN trained on the MS COCO object detection problem. Once we have our data, we’ll use a convolutional neural network (CNN) to classify each frame with one of our labels: ad or football. Stand-alone projects. The main idea of this work is to learn image aesthetic classification and vision-to-language generation using a multi-task framework. There are 61 bands and the size of data file is 1320 x 1040 x 61. In this paper, deep convolutional neural networks are employed to classify hyperspectral images directly in spectral domain. … it is natural to use a CNN as an image "encoder", by first pre-training it for an image classification task and using the last hidden layer as an input to the RNN decoder that generates sentences — Show and Tell: A Neural Image Caption Generator, 2015. One popular toy image classification dataset is the CIFAR-10 dataset. Front Page DeepExplainer MNIST Example¶. I wanted to build on it and show how to do better. Why Multi-Label Classification ? There are many applications where assigning multiple attributes to an image is necessary. The aim of this post is to build a custom Mask R-CNN model that can detect the area of damage on a car (see the image example above). It seems reasonable to use the most promising model for object recognition, which is convolutional neural network (CNN). The set of classes is very diverse. Learning Spatial Regularization with Image-level Supervisions for Multi-label Image Classification intro: CVPR 2017. Tiled Convolutional Neural Networks: Encoding Time Series as Images for Visual Inspection and Classification Using Tiled Convolutional Neural Networks (Z. [email protected]
This is common in machine learning where our scripts are run on some other host with more capabilities. Project Title: Cat vs Dog Image Classifier. Image Classification. ) Develop Fashion Recommendation System; Feature Extraction using CNN; Partner: ZUM Internet; Hangul and Font Recognition Multi-task Learning Model (2017. Description In order to facilitate the study of age and gender recognition, we provide a data set and benchmark of face photos. How is SVM used for classification of images? I have read some articles about CNN and most of them have a simple explanation about Convolution Layer and what it is designed for, but they don. This is a general overview of what a CNN does. Convolutional Neural Networks for CIFAR-10. Tensorflow Image Classification. Converting data from its initial form to a more lowlevel form may improve er. In fact, it is more natural to think of images as belonging to multiple classes rather than a single class. This technique is also often referred to as bag of words. Detection/Segmentation: High-level semantic meaning with high spatial resolution Called U-Net, Hourglass, or Conv-deconv. While sentiment classification is an interesting topic, I wanted to see if it is possible to identify a movie's genre from its description. Images are not guaranteed to be of fixed dimensions and the sign is not necessarily centered in each image. Image-based Fashion Recommendation System (2017. R-CNN at test time. An approach to granite tiles classification using CNN is presented in Ferreira and Giraldi (2017). This page describes common signatures that should be implemented by modules for image-related tasks. Source code for this example is available on François Chollet GitHub. on Computer Vision and Pattern Recognition (CVPR), Boston, 2015. In my post Image classification with pre-trained CNN InceptionV3 I managed to achieve an accuracy of around 80%. This ensures that all of our image intensities fall over the same range - one fewer thing for the network to be biased against. Fast R-CNN replaced the SVM classifier with a softmax layer on top of the CNN to output a classification. • Given an image, classify it as one of the 10 classes • Multi-class classification • Supervised learning • Balanced class Goal: 1) To build different neural network (CNNs, VGG etc. ImageNet classification with deep convolutional neural networks. Very useful for loading into the CNN and assigning one-hot vector class labels using the image naming. Kelp is a mixed breed dog but she has a lot of malamute in her so we'll consider this a success! You can also try with other images, like the bald eagle image:. The two files are both in a format that the C++ and Python image classification examples can read in, so you can start using your new model immediately. In this project, I applied Keras/Tensor Flow to train the neural network model for classification analysis: to predict whether an image is blurry or clear. There are several tricks involved in using CNNs for such a task. Dec 23, 2016. INTRODUCTION There has been significant progress in field of object recognition using deep convolutional neural networks. This article is going to discuss image classification using a deep learning model called Convolutional Neural Network(CNN). Figure 3: Plotted using matplotlib. In other words, a class activation map (CAM) lets us see which regions in the image were relevant to this class. Earlier, we alluded to recent breakthroughs in computer vision using convolutional neural networks. Goal: MRI classification task using CNN (Convolutional Neural Network) Code Dependency: Tensorflow 1. This is mainly due to the number of images we use per class. 1-7, January 05-07, 2018, Langkawi, Malaysia. Videos can be understood as a series of individual images; and therefore, many deep learning practitioners would be quick to treat video classification as performing image classification a total of N times, where N is the total number of frames in a video. For each window the network should output:. Convolutional Sequence to Sequence Learning; Deep Image Retrival. Most major companies such as google  Are using them for help them with their image search tools. Understanding AutoEncoders using Tensorflow; Image Classification using pre-trained models in Keras; Transfer Learning using pre-trained models in Keras; Fine-tuning pre-trained models in Keras; More to come. ) models that can classify given images. Hi, there! I'm a data science researcher, a Master's degree candidate and a writer on Medium 😍. The original Caffe implementation used in the R-CNN papers can be found at GitHub: RCNN, Fast R-CNN, and Faster R-CNN. Description In order to facilitate the study of age and gender recognition, we provide a data set and benchmark of face photos. In this paper we propose cross-modal convolutional neural networks (X-CNNs), a novel biologically inspired type of CNN architectures, treating gradient descent-specialised CNNs as individual units of processing in a larger-scale network topology, while allowing for unconstrained information flow and/or weight sharing between analogous hidden layers of the network---thus generalising the. So is it correct to generate the images like that for the convolutional neural network. A subset of image classification is object detection, where specific instances of objects are identified as belonging to a certain class like animals, cars, or people. The rationale for such a model is that it can be used by insurance companies for faster processing of claims if users can upload pics and they can assess damage from them. The score of each class for an image is then obtained by comparing each template with the image using an inner product (or dot product) one by one to find the one that “fits” best. So even though we're using a single image, we need to add it to a list:. In order to get started with Convolutional Neural Network in Tensorflow, I used the official tutorial as reference. The reason this works well is that the examples in the training data are correlated. On the server side, we will be using high-performance computing GPUs to feed forward the image in a convolutional neural network which is a popular deep learning network. Attention-based Extraction of Structured Information from Street View Imagery. Source: https://github. −Teaching Deep Learning (Slides, code available at uvadlc. ImageNet, which contains 1. Topic-Enhanced Memory Networks for Personalised Point-of-Interest Recommendation arXiv_CV arXiv_CV Attention Relation Memory_Networks Recommendation. Training loss of CNN-Softmax and CNN-SVM on image. In this post we will implement a model similar to Kim Yoon’s Convolutional Neural Networks for Sentence Classification. Building powerful image classification models using very little data. I know there are many blogs about CNN and multi-class classification, but maybe this blog wouldn't be that similar to the other blogs. INTRODUCTION There has been significant progress in field of object recognition using deep convolutional neural networks. We will classify images using a CNN. This blog post is inspired by a Medium post that made use of Tensorflow. Neural Networks in Keras. Abstract: This paper presents a Convolutional Neural Network (CNN) for document image classification. The basics of a CNN architecture consist of 3 components. In this project, I applied Keras/Tensor Flow to train the neural network model for classification analysis: to predict whether an image is blurry or clear. Training loss of CNN-Softmax and CNN-SVM on image. The problem I am working on is Weed classification. The first step is to install the. January 22, 2017. Published in IEEE Workshop on Analysis and Modeling of Faces and Gestures (AMFG), at the IEEE Conf. As per wikipedia, “PyTorch is an open source machine learning library for Python, based on Torch, used for. There are many free courses that can be found on the internet. In order to test my hypothesis, I am going to perform image classification using the fruit images data from kaggle and train a CNN model with four hidden layers: two 2D convolutional layers, one pooling layer and one dense layer. If we detected all valid shelves in a picture, then we could consider that stand as valid. NIMA consists of two models that aim to predict the aesthetic and technical quality of images. 0, Anaconda 4. Among the 10 partitions, 1 partition was used as a test dataset, while all other images were used as training datasets. This is a general overview of what a CNN does. There are 61 bands and the size of data file is 1320 x 1040 x 61. This section contains several examples of how to build models with Ludwig for a variety of tasks. Darknet displays information as it loads the config file and weights, then it classifies the image and prints the top-10 classes for the image. Examples to use pre-trained CNNs for image classification and feature extraction. Classification. More details here: https://arxiv. One concern about LULC classification is that at high resolutions, there is a significant amount of variability in the data. At the end of the article, the reader will be able to develop a simple application which will search into a list of images for the one containing a. For this task I will have to train CNN using pretrained model from ImageNet and apply transfer learning techique. Image Classification with Keras. 33% accuracy). As an output, the module generates a score that indicates. This repository/documentation is done as part of my PhD, if you were to find a mistake or inaccuracy please let me know so I can correct it. Having to train an image-classification model using very little data is a common situation, in this article we review three techniques for tackling this problem including feature extraction and fine tuning from a pretrained network. All gists Back to GitHub. so, if you ever have questions, here were a couple of helpful answers i found:. Examples to implement CNN in Keras. Deep Learning is a new area of Machine Learning research, which has been introduced with the objective of moving Machine Learning closer to one of its original goals: Artificial Intelligence. Why do we use it then?. GitHub Gist: instantly share code, notes, and snippets. TensorFlow is an open-source machine learning library for research and production. Converting data from its initial form to a more lowlevel form may improve er. FastGFile (image_path, 'rb'). There are several tricks involved in using CNNs for such a task. Transfer Learning using pre-trained models in Keras; Fine-tuning pre-trained models in Keras; More to come. Using CNN to classify images in KERAS Let’s try taking an example where we input several images of cats and dogs and we try to classify these images into their respective animal category. Published: Exploiting Image-trained CNN Architectures for Unconstrained Video Classification Efficient Video Classification Using Fewer. A brief introduction to CNNs is given and a helper class for building CNNs in Python and TensorFlow is provided. Reference¶ Deep Inside Convolutional Networks: Visualising Image Classification Models and Saliency Maps; keras-vis; Reference in this. Implementing a CNN for Text Classification in TensorFlow The full code is available on Github. import tensorflow as tf # change this as you see fit image_path = sys. That itself would be easy to do using image processing. Task: look at the Figure 1 generated by the code and at the code itself and make sure that you understand what you are looking at. Chellapilla et al. January 21, 2017. Google Neural Machine Translation System. Load the pretrained AlexNet neural network. The main purpose of this project is to create a web applications for quick image to image translation. Train a CNN model (similar to AlexNet) on the image classification task. Image-based Fashion Recommendation System (2017. ImageNet Classification with Deep Convolutional Neural Networks deep convolutional neural network to classify the 1. The aim of this post is to build a custom Mask R-CNN model that can detect the area of damage on a car (see the image example above). Biological Connection. eral deep learning models on the image classi cation problem. My second takeaway is that there is a lot of work to be done here. By continuing to use this website, you agree to their use. Since this project is going to use CNN for the classification tasks, the original row vector is not appropriate. Just as its literal meaning implies, a feature is a distinct and useful observation or pattern obtained from the input data that aids in performing the desired image analysis. Classifying “out in the wild” moving images with 101 classes while only looking at individual images at 65% accuracy is just astounding. cc/paper/4824-imagenet-classification-with. But, in this post, I have provided you with the steps, tools and concepts needed to solve an image classification problem. # Grab an image from the test dataset img = test_images print(img. First part of image description may be solved by detection task. I will be using classical cat/dog classification example described in François Chollet book — Deep Learning with Python. CNNs for Classification: Class Probabilities •CONV and POOL layers output high-level features of input •Fully connected layer uses these features for classifying input image •Express output as probability of image belonging to a particular class 88. It seems reasonable to use the most promising model for object recognition, which is convolutional neural network (CNN). California State University. Kelp is a mixed breed dog but she has a lot of malamute in her so we'll consider this a success! You can also try with other images, like the bald eagle image:. Training accuracy of CNN-Softmax and CNN-SVM on image classification using MNIST. Road Damage Detection And Classification In Smartphone Captured Images Using Mask R-CNN classification of damage in state-of-the-art GitHub badges and help. Neural Networks in Keras. Examples to use pre-trained CNNs for image classification and feature extraction. Image classification using CNN. As per wikipedia, "PyTorch is an open source machine learning library for Python, based on Torch, used for. 8, Python 2. You can refer to the attached github project, which works on video classification. , a deep learning model that can recognize if Santa Claus is in an image or not): Part 1: Deep learning + Google Images for training data. Image Recognition ( a. of image set clustering for object classi cation as well as ne-grained classi cation. Text classification using CNN. imdb_cnn_lstm: Trains a convolutional stack followed by a recurrent stack network on the IMDB sentiment classification task. The path is where we define the image location and finally the test_single_image cell block will print out the final result, depending on the prediction from the second cell block. The shape of X_test is (10000, 28, 28). GitHub has ≈6M of active users (and 3M after some filtering). Fine tuning of a image classification model. https://github. handong1587's blog. Image Classification with CNNs. An approach to granite tiles classification using CNN is presented in Ferreira and Giraldi (2017). Below you can see an example of Image Classification. Just with any problem in Artificial Intelligence, we first need to formulate our problem. Dense Adversary Generation The Dense Adversary Generation (DAG)  approach. You can consume some of the multiple pre-trained models, build your own model or even use a Keras Model! So yes, we need a microservice for image classification. It got quite popular. This method aims to allow using the high-resolution histopathological images from BreaKHis as input to existing CNN, avoiding adaptations of the model that can lead to a more complex and. Background. Approaches to developers clustering Approaches to developers. Language and vision are processed as two different modal in current work for image captioning. This blog is based on the tensorflow code given in wildml blog. Image Classification with CNNs. The pipeline shows heterogeneous scenarios that use an IPU and GPU in parallel to the CPU, and advanced concepts such as a tiled user using custom OpenCL kernels. A separate category is for separate projects. The dataset we'll be using in today's Keras multi-label classification tutorial is meant to mimic Switaj's question at the top of this post (although slightly simplified for the sake of the blog post). There are 61 bands and the size of data file is 1320 x 1040 x 61. Finally, viewpoint-invariant age and gender classiﬁcation was pre-. Classification. Researchers and enthusiasts alike, work on numerous aspects of…. Three methods of training using a Conv Net: Supervised. The CIFAR-10 small photo classification problem is a standard dataset used in computer vision and deep learning. In the past, I have written and taught quite a bit about image classification with Keras (e. ) models that can classify given images. This is mainly due to the number of images we use per class. Hence, we aim at an improvement of surgical action classification by using early fusion of raw image data with multiple (domain-specific) modalities of which at least one represents a temporal dimension, such as motion vectors. We found out that spectrogram image classification with CNN algorithm works as well as the SVM system. One popular toy image classification dataset is the CIFAR-10 dataset. The path is where we define the image location and finally the test_single_image cell block will print out the final result, depending on the prediction from the second cell block. ipynb - trains and evaluates the Neural Network using a feed dictionary. These notes accompany the Stanford CS class CS231n: Convolutional Neural Networks for Visual Recognition. Five video classification methods implemented in Keras and TensorFlow to continuously pass our images through the CNN every time we read the same sample or train. Watch Queue Queue. You can even use Convolutional Neural Nets (CNNs) for text classification. Chellapilla et al. January 21, 2017. It shows how to use layers to build a convolutional neural network model to recognize the handwritten digits in the MNIST data set. Mask R-CNN - 용어 정리를 중심으로 정리 Background Knowledge. on Computer Vision and Pattern Recognition (CVPR), Boston, 2015. Then, we replace the top classifier layers by a regression network and train it to predict object bounding boxes at each spatial location and scale. A similar situation arises in image classification, where manually engineered features (obtained by applying a number of filters) could be used in classification algorithms. Hence, it is now my goal to achieve similar accuracy with the dog breed dataset, that has much more categories, while it is much much smaller. Decoder_Spatial_CNN. They’re used in practice today in facial recognition, self driving cars, and detecting whether an object is a hot-dog. multiclass classification using tensorflow. , image classification modules tend to do some feature extraction on the way). Abstract: This paper presents a Convolutional Neural Network (CNN) for document image classification. It is where a model is able to identify the objects in images. FastGFile (image_path, 'rb'). Current approach: Run K-means with K=45000 (using src-d/kmcuda) Run t-SNE to visualize the landscape; BTW, kmcuda implements Yinyang k-means. I ignored first several layers, and used only features from layers 6-10; There are 1050 feature maps in toatl on this layers, so each image is described with a binary vector of length 1050. This blog is based on the tensorflow code given in wildml blog. Real-Time Document Image Classification using Deep CNN and Extreme Learning Machines for document image classification. This paper is concerned with a new approach to the development of plant disease recognition model, based on leaf image classification, by the use of deep convolutional. Using CNN to classify images in KERAS Let’s try taking an example where we input several images of cats and dogs and we try to classify these images into their respective animal category. document classification, or document segmentation. After using convolution layers to extract the spatial features of an image, we apply fully connected layers for the final classification. 0, Anaconda 4. In this tutorial, we will discuss how to use those models as a Feature Extractor and train a new model for a. By using the AlignDlib utility from the OpenFace project this is straightforward:. One interesting follow-up question for future investigation is how computer-level accuracy compares with human-level accuracy on more complex image understanding tasks. ) models that can classify given images. preprocessing. NET wrapper for the Intel OpenCV image-processing library. MNIST 데이터는 grayscale 이미지기 떄문에 3번째차원(컬러채널)의 값은 1입니다. intro: “propose an architecture consisting of a character sequence CNN and an N-gram encoding CNN which act on an input image in parallel and whose outputs are utilized along with a CRF model to recognize the text content present within the image. Jiyang Gao, Chen Sun and Ram Nevatia, “ACD: Action Concept Discovery from Image-Sentence Corpora” in ACM International Conference on Multimedia Retrieval (ACM ICMR), 2016 [Oral], arxiv. I’ll be showing you how to train your CNN in today’s post using Keras and deep learning. Create a label_image. ImageNet Classification with Deep Convolutional Neural Networks. ImageNet VGG16 Model with Keras - Explain the classic VGG16 convolutional nerual network's predictions for an image. of Electrics & Electronics Engineering Nanyang Technological University. Here, we use Dlib for face detection and OpenCV for image transformation and cropping to produce aligned 96x96 RGB face images. The well-known application of CNN is image classification, where a fixed dimension image is fed into a network along with different channels (RGB in the case of a color image) and after various steps of convolution, pooling and fully connected layers, network outputs class probabilities for the image. The set of classes is very diverse. INTRODUCTION There has been significant progress in field of object recognition using deep convolutional neural networks.