Questions tagged [computer-vision]

Questions related to image representation, segmentation, visual object categorization and image processing algorithms in general.

Filter by
Sorted by
Tagged with
0
votes
0answers
8 views

Semantic Segmentation Multi-Class Single Channel Output Math

For semantic segmentation problems, I understand that it's a pixel-wise classification problem. At the last layer of the neural network, I would basically have a 1x1x1 convolution layer with a softmax ...
1
vote
1answer
21 views

Performance of MaskRCNN/YOLO as a function of object size in pixels

I am trying to find references on how the resolution of an object affects the ability of object detection systems such as MaskRCNN and YOLO to correctly identify the object. For example, if the ...
1
vote
1answer
24 views

Incorrect predictions on extracted images from text

I trained a model in PyTorch on the EMNIST data set - and got about 85% accuracy on the test set. Now, I have an image of handwritten text from which I have extracted individual letters, but I'm ...
0
votes
0answers
12 views

video normalization using skvideo

I'm building a model that would take as input a video of 25 frames and would (ideally) output the next 25 frames. My question is when we use images we usually normalize by dividing the X by 255. ...
0
votes
0answers
11 views

Image classifier from text files

I have a data set of thousands of images of hundreds of pixels in gray scale ranging from -1 to 1. The labels represent 0 to 9. The issue is that the data set is in .txt format. How can one make an ...
3
votes
1answer
43 views

Same value of min and max in min-max normalisation

By the definition of min-max normalisation, the value is divided by max - min, what if the max and ...
1
vote
0answers
22 views

Karush-Kuhn-Tucker Conditions in a Biometrics Research Paper

For one of my course works I'm supposed to find a scientific paper from my field of research that involves an optimisation problem solution based on the Karush-Kuhn-Tucker conditions. My research ...
2
votes
1answer
70 views

What would be the ideal dataset to train a model to detect advertisements in an image?

I am thinking of the requirements for training a model that would be able to detect if there is any kind of ad in an image. I know that this sound too broad not just for a question on CV but for ...
2
votes
3answers
157 views

Which image format is better for machine learning .png .jpg or other?

I'm trying to train a neural network with images. Since I'm extracting images from a video feed I can convert them either to .png or .jpg. Which format is preferred for machine learning and deep ...
1
vote
0answers
19 views

Reconstruction using low-light images

Let's say we have a regular photo and three low-light photos illuminated in different colors. Each pixel is a three-component vector $q=(R,G,B)$. Then $q_k^{A}$ is the $k$-th pixel of the regular ...
0
votes
1answer
32 views

Understanding the weighted cross-entropy method of u-net

I am trying to implement the weight-cross entropy mentioned in unet paper to counter the class-imbalances. I am not really able to understand how they are exactly implementing the weight-cross entropy....
0
votes
1answer
26 views

Object detection : Multimodal or single input ? for Depth + Thermal images

I need to detect persons in a scene. I have a 16 bits depth image of that scene (640, 480) I have a 16 bits thermal image (80, 60) of the same scene (slighly different point of view) I resize these ...
1
vote
0answers
13 views

Understanding the concept behind Microsoft Computer Vision- Analyze an Image. Also extract metadata of image from different search engines

My task is to upload an image and get as much information as I can get about the image or in different words its metadata (not exif) like - what is it, caption, price, what does it do, application, ...
0
votes
1answer
22 views

Semantic segmentation mask

What does the mask look like when doing semantic segmentation. I have 3 classes (background, liver, tumour). Currently the input to my segmentation model looks like this (32, 128, 128, 3) where 32 =...
0
votes
1answer
33 views

Object classification

I'm currently working on a "Where's Waldo" project as part of my coursework, where I have to find 3 different characters in any given image - Waldo, Wenda, and Wizard. I'm trying to convert this ...
1
vote
0answers
20 views

Shape classification by color - computer vision

I have task of classification of polygons. These shapes may differ only by color, or be transformed by linear transformation (scaled, rotated). I am using ORB for separating different polygons, but it ...
1
vote
1answer
30 views

Object detection with just one image? [duplicate]

Deep neural networks require lots of examples to learn tasks like image classification, and object recognition. On the other hand, we humans can learn and identify object just by looking at it once. ...
0
votes
1answer
49 views

batch size of stochastic gradient descent [duplicate]

I understand that stochastic gradient descent has a batch size of 1, but while reading inception v2 paper, I found this text in training methodology "We have trained our networks with stochastic ...
0
votes
1answer
15 views

Is there a difference between training with multiple objects in a single image and multiple objects in a different images?

I'm trying to generate data for my object detection network (which will be used for TensorFlow: ResNet). What I'm currently curious about is this: if I have the same total amount of data (each data ...
0
votes
0answers
31 views

Is there a model or algorithm to improve digital drawings?

given a bad drawing, the algorithm should deform the edges of the bad drawing and fill them with color so that it looks more like a learned character. the entry would be a very poorly made drawing of ...
0
votes
0answers
14 views

Impact of increasing number of classes on object detection rate/speed

I am trying to build a Darknet (C++) model to detect custom objects. Initially, in the testing phase for 50 objects with tiny yolo, I am getting 35 FPS speed. If I increase the number of objects to ...
0
votes
1answer
49 views

Is it possible to train a model to detect a shape? [closed]

We have multiple object detection APIs that can help us to find the bounding box coordinates. But is it possible to go a little further and find/separate the shape of the object (say cat/dog/human) ...
0
votes
0answers
21 views

What do I do when my deep learning model in CNN does not learn? [duplicate]

I'm training on a skewed image dataset of 5000 images with class weights. The training loss decreases well and swift, but the validation-set loss fixates after around 6 epochs. I have tried using my ...
0
votes
1answer
66 views

Object distortion after ROI Align in Mask R-CNN

In Mask R-CNN, if there are 2 proposed ROIs which cover 2 objects that looks like below: #1 A square object #2 A rectangular object So my question is: After ROI Align, is the #2 feature map ...
2
votes
1answer
24 views

Mode-collapse problem in GAN: How to grasp the full entropy of the distribution we want to model in GAN

In pix2pix GAN paper( https://arxiv.org/abs/1611.07004), authors found that the noise vector and the dropout are not efficient in grasping the full entropy of the data distribution we want to model. ...
0
votes
0answers
7 views

Model for simultaneous person detection and pose estimation

Does someone know a model that performs person detection (eg using a bounding box like YOLO or Mask-RCNN) and simultaneously pose estimation (like CPM or Personlab) in one forward pass. The models I ...
3
votes
1answer
109 views

Training an Object Detection Model Using with Artificial Data from Video Games

I had an interesting idea of using artificial data gathered from screen shots of a high-resolution video game as a cheap substitute for labeled real data, which can be quite expensive or difficult to ...
0
votes
0answers
60 views

Training data for extracted license plates from car images

I am working on a project which uses machine learning and image processing techniques to detect/extract license plates of a vehicle given an image. In my module for data preparation and feature ...
2
votes
0answers
29 views

How to use machine learning to create combine of opposite images side by side [closed]

Inspired by: Two Worlds Pictures I just want to create a Machine Learning Model that can automatically combine the opposite images into 1 image. I am thinking about 2 possible solutions: Pose ...
1
vote
1answer
28 views

Does the model architecture of a CNN depend on the dimension of your input images?

By model architecture, I'm interested in knowing the following: Number of nodes in input layer Number of nodes in subsequent layers Number of layers in the architecture Number of filters and ...
0
votes
0answers
38 views

Chinese character recognition from generated images - Validation accuracy does not improve

I am currently working on creating a simple Chinese character recognition network. Given an grayscale image of a character, the goal is to predict the depicted character. I want to run the model on a ...
1
vote
0answers
8 views

Finding objects in an entanglement

I'd like to know if there are ways to solve entanglement problems like this one: Given a known shape (in red), can a machine locate (or at least count) the instances of this same object in an image ...
0
votes
1answer
63 views

How do convolutional neural networks learn from images of different translations and conditions?

when we feed the CNN images of cats in different lighting conditions or colors, Is it the job of the conv layers to learn the different representations(lighting conditions and colors) and map them to ...
0
votes
1answer
24 views

How to correctly find the best hyperparameter combination when training a neural network?

I am not sure whether this is the right place to ask this question, so feel free to redirect me if not. What I'm doing is bench-marking a model (MobileNet v2 100 224) in terms of performance - size ...
0
votes
0answers
92 views

How to reproduce the classification heatmap figure (Figure 2) from the FCN paper?

I want to replicate the results of the paper "Fully Convolutional Networks for Semantic Segmentation" (Long and Shelhamer, 2015) in PyTorch. Doing so, I tried to replicate the initial experiment ...
1
vote
1answer
402 views

100% accuracy on training, high accuracy on testing as well. What does this mean?

I was training a model to classify different traffic signs and decided to use a pre-trained alexnet model and redefining the last fully-connected layer to match the classes of the dataset. When I did ...
0
votes
0answers
38 views

When should I stop the object detection model training while mAP are not stable?

I am re-training the SSD MobileNet with 900 images from the Berkeley Deep Drive dataset, and eval towards 100 images from that dataset. The problem is that after ...
0
votes
0answers
21 views

Tune the input size of the image on known architecture

Known CNN and FCN architecture (in computer vision), such as Inception, Resnet, Alexnet, etc, have a specific input size, that can't be changed once the architecture has been chosen. Everyone that ...
2
votes
2answers
266 views

Object detection : is deep learning the only way to go?

It seems that deep learning based approaches are currently more superior to the more "traditional" methods in the domain of object detection. Methods like YOLO, for example, seem to be doing something ...
0
votes
0answers
22 views

Which model should I choose for object detection and classification

My use case is to detect the defects in vegetables in an isolated system with high accuracy and speed. Currently, I'm using Tensorflow Object Detection API (Faster RCNN) for this purpose. But I've ...
1
vote
1answer
100 views

Why are we interested in top-n accuracy?

I know the definition of top-n accuracy: [1] What is the definition of Top-n accuracy? My question is, why do we even care for this? A short answer could be to compare different models in various ...
0
votes
0answers
10 views

What are some neural network models that can use auxiliary info during training for image segmentation?

What are some deep learning models that can use supplementary information other than RGB channels for image segmentation? For example imagine a poorly shot image of a river (blue) that shows a gap, ...
5
votes
2answers
130 views

Reference request - Computer Vision Book

What are the best books for obtaining a strong understanding of computer vision? From what I understand based on my undergraduate class, almost all current state-of-the-art computer vision is just ...
0
votes
1answer
59 views

How to deal with different scales with CNN?

In the deep learning book by Goodfellow et al., it is stated "Convolution is not naturally equivariant to some other transformations, such as changes in the scale or rotation of an image. Other ...
1
vote
1answer
14 views

Same input size for Source and Target Model?

For the transfer learning do we need to have same input image size of target model as source model? for example my source model is trained on 100x100 input images and target model is low resolution ...
0
votes
0answers
10 views

What types of problems can object detection not be used on?

I'm curious what types of problems can object detection(ex: Faster-RCNN) not be used? For example, I'm guessing if you were trying to detect an object which is conditioned on something else occurring ...
0
votes
0answers
22 views

YOLOv3 Detection Layer

I'm learning about YOLOv3 and wanted to confirm my understanding of how the detection layer works. Just to confirm, we have some output tensor from which we would filter for the objectness score above ...
0
votes
1answer
22 views

How to deal with different image sizes during training and inference? (e.g. Stacked Hourglass)

In some of computer vision papers I read that they start off with a bigger sized image and use pooling to reduce dimensionality and train on the image with lower resolution. However, they don't ...
0
votes
0answers
154 views

What is the best way to normalise image data?

The normalisation in an image really confuses me. I mean there are multiple ways to do it (see below) but, is there the best one, or most preferable one, or one needs to experiment with all to find ...
0
votes
0answers
37 views

Unsupervised classification of images

Assuming I have a dataset of images from two similar classes, for example let's say 95% Labradors and 5% of Chihuahuas and I want to make a classifier. The point is that I need to find the anomalies (...