Stack Exchange Network

Stack Exchange network consists of 175 Q&A communities including Stack Overflow, the largest, most trusted online community for developers to learn, share their knowledge, and build their careers.

Visit Stack Exchange

Questions tagged [cnn]

Convolutional Neural Networks (CNN, also called ConvNets) are a tool used for classification tasks and image recognition. The name giving first step is the extraction of features from the input data.

1
vote
0answers
8 views

Time depending images as input in a CNN

I'll try my best to be understood, thank you for reading. So this is the problem : I'm asked to predict cancer on X-ray as early as possible. The dataset is composed of images in this form ...
0
votes
0answers
15 views

CNN validation accuracy not improving - spectrogram

I am new to Machine Learning. So, for a project I am trying to classify instruments in .wav file. The dataset I am using is IRMAS. Dataset contains 11 classes of instruments with recordings in 16 bit ...
1
vote
0answers
10 views

Configuration of masks in PixelCNN

I have trouble understanding the masks in PixelCNN 1. Here for A, each channel is connected to prior channels. For B, channels are also connected to themselves. My problem is in first layer when I ...
2
votes
1answer
35 views

Why does my minimal CNN example show strongly fluctuating validation loss?

I'm fairly new at working with neural networks and think I am making some basic mistake. I am trying to assign simulated images to 5 classes to test, what (if any) networks are helpful for a large ...
0
votes
0answers
12 views

Convolution Neural Networks on microcontrollers

I am wondering if an already trained convolutional neural network can be represented as a formula just like a perceptron can ($x_1w_1 + x_2w_2 + \ldots =$ PREDICTION). I know that the formula could be ...
0
votes
1answer
14 views

Does Convolution kernel size affect number of channels?

I am going through Dilated Residual Network blog post. In this, Under 2.Multi-scale Context aggregation heading, author mentioned this. The last one is the 1×1 ...
2
votes
1answer
20 views

How to visualise GIST features of an image

I am currently working on a image classification application using deep learning algorithms (either by using GIST features or CNN). I need help in understanding the below queries. I have extracted ...
0
votes
1answer
18 views

Character Fonts Classification - low accuracy [on hold]

I am trying to develop a multi-class classification model with keras. The task is similar to MNIST but applied to this character font image dataset. I have 153 classes and around 900k entries. I am ...
0
votes
1answer
18 views

Model for classifying time-series data with distinct features?

I've heard about time-series classification being done with TCN's and CNN's combined with LSTM's very often, citing that CNN's would provide insight both forward and in the past since you already have ...
1
vote
0answers
12 views

Convolution Neural Network with 1 input channel

I am trying to make a sudoku solver using a convolution neural network with pytorch. I am failing to implement the architecture with one channel. There are many great tutorials which deal with images ...
1
vote
0answers
13 views

What is the input for Conv2d for text?

I have an embedding layer that returns a tensor of type: batch x max_len x embeddings_dim (8 x 200 x 300) I want to input this to a Conv2d layer with kernel 3x3. I tried ...
1
vote
0answers
43 views

Precision decreases with each epoch in CNN

I've built the following CNN that is used to classify a binary classification set (something a like cats or dogs): ...
1
vote
0answers
12 views

Weight update to fully convolutional network when supervision is only for a patch

I had a fundamental question, that is independent of any DL framework. In a fully convolutional network, if we forward an image of size 1000 x 1000, but only provide supervision signal for a 100 x 100 ...
3
votes
1answer
42 views

CNN - imbalanced classes, class weights vs data augmentation

I have a set of data with a few strongly imbalanced classes, eg the smallest class is about 54 times smaller than the largest. Therefore, data augmentation in order to equalize the size of classes ...
3
votes
1answer
20 views

How to choose the number of output channels in a convolutional layer?

I'm following a pytorch tutorial where for a tensor of shape [8,3,32,32], where 8 is the batch size, 3 the number of channels and 32 x 32, the pixel size, they define the first convolutional layer as ...
3
votes
0answers
49 views
+50

convLSTM : how to structure input data

I have the following dataframe containing training data that I have been using to perform a regression task using CNN + FC : ...
0
votes
1answer
19 views

review: gradient descent, epochs, validation in neural network training

These days, training data aren't put in gradient descent all at once. Rather, they are put in batch after batch. Gradient descent is run once for each batch of training data. When all batches are ...
0
votes
0answers
15 views

Remedies to CNN-LSTM overfitting on relatively small image dataset

Notes Using a pretrained model, trying data augmentation (not possible knowing nature of images, lowering number of parameters in the network, all didn't help) Context I have a sequence of images. ...
0
votes
1answer
35 views

What happened to the accuracy before and after 75th epoch. Why its unstable at first, then stepped up after the 75th epoch?

What happened to the accuracy before and after 75th? Why before 75th, its unstable and suddenly stepped up after that? I am using RESNET and CIFAR-10 dataset.
-1
votes
0answers
17 views

What's a right way to approach such an image classification project?

Please advise how would you approach development of ML for such a project? http://vps389544.ovh.net:5555/. It's trained using this dataset tamaraberg.com/street2shop and works based on CNN as far as I ...
2
votes
1answer
29 views

What's the right way to setup an image classifier by multiple params?

I'm very new to the data science and machine learning, so apologies for my ignorance. What I'm trying to understand is how to setup an image classifier system (maybe based on CNN) which will classify ...
0
votes
0answers
6 views

How can I enrich train data in case of cnn using target and time features

I have a sequence of images, let's say we ignore time specificity for now. In the other hand, target is a multivariate continuous time series. Let's consider it just a univariate one. Training a cnn ...
0
votes
0answers
10 views

Image classification using Semantic Segmented Images

Can we use the semantic segmented images directly to perform image classification using CNN model? Updating the question: I am trying to classify images as below: a. Input : Images taken from camera ...
0
votes
0answers
50 views

Model loss and validation loss not decreasing? How to speed?

Every image was resized to 256x256 pixels. Batch size = 4. (GPU GTX 1050 memory ~4GB). The mask R-CNN model was initial-ized using pretrained weights from COCO dataset TRAIN_ROIS_PER_IMAGE = 8 ...
0
votes
1answer
7 views

Choose CNN architecture first, then optimize parameters - validation vs test performance to pick architecture?

I am doing a few experiments on medical data. I am about to transfer learn the pretrained networks for my problem. Firstly, I have to pick a network architecture. Secondly, I would like to optimize ...
2
votes
1answer
58 views

Identify credit card shape using machine learning

I have to approach this task : identify credit card from a image. I am attaching example image below I have to identify and localize credit card from this image. The real challenge is that the card ...
1
vote
1answer
25 views

Optimization based on validation and not training

Hello neural network programmers, I am currently creating a neural network with keras, as I am not that familiar with tensorflow and it's a bit more difficult. I want my optimization to optimize the ...
0
votes
0answers
16 views

How does Max Pooling (of size 2x2) change the size of receptive field in CNN?

I am often told that Max Pooling of 2x2 doubles the size of the receptive field from the previous layer. Would like to understand if that is true, how that happens. I have already checked this ...
0
votes
0answers
13 views

Encoder Decoder Network Image Compression

Could you train an encoder decoder network to take an image in and attempt to recreate that image as an output. I am basically interested at looking at the intermediate feature vector representation ...
0
votes
0answers
9 views

What (and how) should I use for object detection in this case?

I have read a few papers on object detection and analyzed several implementations of object detection applied to car number plates. In my case, the task is a bit more difficult in that I need to ...
0
votes
1answer
56 views

Need equations for some of weight initializers in tensorflow?

For my CNN model I tried some of the weight initializers such as "truncated normal initializer","random normal initializer","glorot normal initializer","glorot uniform initializer", my questions are: ...
3
votes
1answer
61 views

The most used loss function in tensorflow for a binary classification?

I am working on a binary classification problem using CNN model, the model designed using tensorflow framework, in most GitHub projects that I saw, they use "softmax cross entropy with logits" v1 and ...
1
vote
1answer
21 views

confused about parameter updates and forward/backward pass according to batches and epochs in CNN?

I am working on a CNN model, the code written in tensorflow, I did some googling about parameter updates such as weights ana biases when method is optimized and the loss is computed, two things made ...
1
vote
1answer
35 views

Understanding of Convolutional Neural Network (CNN) [closed]

I am a beginner in ML and I would like to learn CNN with Math behind. If you suggest any good blog or documentation which help me get depth knowledge about CNN.
1
vote
0answers
15 views

How to use pre-trained weights to initialize the custom CNN?

From this paperhere, it shows that U_Net initialized by VGG received a better result than the one trained from scratch. Now I want to build a custom u_net which has [32,64,128,256] which is different ...
1
vote
1answer
21 views

Channels in convolutional layer

I usually see convolutions performed over all the channels of the input. For example a $3x3$ kernel is really a $3x3xN$ kernel for a an input with $N$ channels, thus resulting in a single output ...
0
votes
0answers
19 views

Applying CNN for cross sectional data

I have a data-set with around 1K features along with 200K observations and a target having 5 different categories. My target is amount categories containing how much a client can invest (e.g. 0-25K ...
2
votes
1answer
24 views

Papers on anger detection in dialogues

I am interested in anger detection in dialogues and I want to study multiple methods like LSTM, CNN, etc. Are there any good research papers or books about this subject?
2
votes
1answer
22 views

How should continuous outputs be in convolutional neural network?

I have labeled faces images (label is the age -continuous value-) dataset and I want to construct a Convolutional Neural Network model to predict the age of a person. I have the following questions. ...
1
vote
1answer
29 views

Meaning of equation for CNN probabilities

So the first equation above refers to a CNN (rather a committee of CNNs) for image classification. I am unable to understand exactly what the author is trying to do in the first equation. So far, I ...
1
vote
1answer
26 views

What are the possible values of a filter in a CNN?

I am a trying to write a CNN from scratch in python but I am bit new to CNNs specifically the convolution layers as I am comfortable with the dense layers. I was reading Do filters have different ...
1
vote
1answer
22 views

How to classify nationality name on the bases of user driving license with tensor-flow keras python?

I am new to deep learning. I want to create a classifier which can predict nationality names on the bases of nationality on driving license id. To accomplish that, I created a data set of USA driving ...
4
votes
1answer
116 views

How to detect blocks of texts in document images

I am planning to detect texts from document text images like below: GOAL: WORK DONE: I have tried to solve this with some scene text detection algorithms like EAST Text detector and PixelLink. But ...
0
votes
0answers
12 views

Transfer learning on a CNN for computer vision problems

Can a pretrained CNN for semantic segmentation be used for image classification, given the dataset is similar for both ? I am thinking that answer should be 'yes'. But when I tried replacing decoder ...
2
votes
0answers
22 views

What is the motivation for row-wise convolution and folding in Kalchbrenner et al. (2014)?

I was reading the paper by Kalchbrenner et al. titled A Convolutional Neural Network for Modelling Sentences and am struggling to understand their definition of convolutional layer. First, let's take ...
1
vote
0answers
21 views

Image preprocessing: How to resize / align / cut images of various sizes?

I want to create a new dataset for image recognition. If I have Object A that I want to recognize in images, and multiple images of various different sizes, how do I preprocess them so that they ...
0
votes
0answers
4 views

Is it good in general to subtract background from a sequence of images for learning?

Context: I have a sequence of satellite images, indexed by time, so basically it's a video. Images were taken on top of a mountain, to capture a main cause that affects solar rays. (GHI in other ...
0
votes
1answer
14 views

Splitting a neural network in 2 microservices

I have a neural network, that is already trained locally that can detect objects in the scene. But I have to split the neural network into 2 parts, let's say it has 16 layers, and I want to have one ...
1
vote
1answer
20 views

Dataset Made of Multiple Datasets

I recall a tweet by Andrej Karpathy sometime a while ago about a ML dataset made up of cifar-10, mnist, ImageNet, and maybe others. I want to use that dataset, but I can't find it. Does anyone know if ...
4
votes
1answer
36 views

CNN computing time on good CPU vs cheap GPU

I am a researcher working on my first deep learning project, which consists of using a CNN (pre-trained VGG16+2 densely connected layers) to classify drone imagery of vegetation. In trying to hack ...