Questions tagged [computer-vision]

Computer Vision is a subfield of computer science which deals with analyzing and understanding images. This includes detection of objects like faces in images or segmenting images.

Filter by
Sorted by
Tagged with
1
vote
1answer
22 views

Normalization vs standardization for image classification problem

For day and night image classification, is it better to normalize or standardize images? In general, when should I use each method? I am interested in with example why one method is preferred over ...
0
votes
0answers
15 views

crop part of image from image, change it and put it same place in original image

I have image (think of image of small plant) with bounding box. For data augmentation I want to rotate (45 degrees) the image (plant) in bounding box, and place it at same place in original image. so ...
0
votes
0answers
9 views

Correct dimensions of a Siamese network Input array

I have an image dataset where the folder structure is as follows- there are 900 folders (all of which will be classes) and in each folder, we have a varying number ...
0
votes
0answers
8 views

Unet Overfitting for binary segmentation of fake images

I am working on a project where I am trying to detect and localize forgeries in images. I am using the CASIA v2 dataset and using Unet model for the task. I have the binary masks of all the images in ...
0
votes
1answer
10 views

Training the network with some batch size - code

There is my "training" code below, I wrote it based on one youtube tutorial. I don't understand actually one part: batch_X = train_X[i:i+BATCH_SIZE], batch_y = train_y[i:i+BATCH_SIZE]. How ...
1
vote
0answers
11 views

Why YOLO algorithm predicts B boxes for each grid cell S?

In yolo each grid cell predicts multiple bounding boxes lets say in YOLOv1 it predicts B=2, what is the advantage as it only predicts class probabilities only once for each grid cell. If that so why ...
1
vote
1answer
18 views

One hot vector output in classification task

I'm working on CNN model and I used one hot vector type of labels. The number of classes is 3: [1,0,0], [0,1,0], [0,0,1]. net(x) I'm getting such an output: [0....
0
votes
1answer
10 views

Checking trained CNN on the images

I trained my CNN (model) classifier and want to check it on some new images. I have image x, so this syntax works for me for one image: torch.argmax(model(x)) What ...
0
votes
0answers
11 views

Using machine learning to find the most similar image that contains another image

As the title states I want to use ml (maybe some kind of CNN autoencoder?) to find the most similar image (I have a list of 10k+ images) within another image. I am currently just using opencv with ...
0
votes
0answers
14 views

computer vision or machine learning disasters [closed]

overfitted models can often have terrible outcomes. We saw one news where a woman was killed by a self driving car. However, that was the only exam I have read about. Have there been any other major ...
0
votes
0answers
29 views

CNN Color Invariance

For the task of detecting an object of any color, if we train our CNN only with images that only contain that object in one or two colors, will the accuracy of our model's predictions be affected for ...
0
votes
0answers
6 views

Small object detection preprocessing

I am currently working on an object detection project in which I amtrying to detect very small objects 50x50 object in a 2k image. EfficientDet produced a very low result if I just put the raw ...
0
votes
0answers
12 views

How to annotate the object more efficiently

I want to perform object detection and object counting on the given below image using neural networks. My first step will be to annotate and label each object present in the training set of images. In ...
0
votes
0answers
11 views

Value of AP and AR are -1.000 in evaluation

As I understand, the AP amd AR calculation are as of follow: ...
0
votes
0answers
24 views

Testing trained model on the image from the test set

I trained my EfficientNet (CNN) and got accuracy=0.73. The question is how to check it on one concrete image from the testing set? How to write a code in python for it? I described the testing set ...
0
votes
1answer
21 views

What is better opensource alternative for identifying small face other than yolo?

I was trying to identify small face meaning that I want to know who that face belong to according to training dataset. I have previously use yolov4 to detect small object before and I know the ...
0
votes
0answers
7 views

Does PIFUHD 3D reconstruction model preserve sizes?

This is little conceptual question. I am working on 3D reconstruction using deep learning from 2D images. I came across PIFUHD model which is developed by Facebook and I have a question regarding that ...
0
votes
0answers
3 views

Train only Region Proposal Network in faster RCNN architecture

I am looking for a way to used my pretrained EfficientNetv2 model and turn it into an object detection. Is there anyway, I can put my pretrained model as a backbone and only train the region proposal ...
0
votes
0answers
14 views

Vgg16 model validation accuracy is stuck

I am working on a CNN model for MRI brain images classification (Alzheimer disease), I use transfer learning method for image classification - vgg16 model trained on ImageNet (1000 classes). I’ve ...
0
votes
1answer
27 views

CNN can't predict images outside the dataset

I am using celeba dataset to train my CNN face landmark detection model. Here is my model ...
0
votes
0answers
7 views

How can I use COCOAPI/PyCOCOTools to evaluate the performance of Tensorflow Lite models

I have used the Tensorflow Object Detection API to train models on a custom dataset. The tensorflow object detection API also allows evaluating the trained models on a test set and gives results in ...
1
vote
1answer
20 views

face landmark detection cnn loss not converging in tensorflow

I am trying to build face landmark detection model using simple regression.I used celeba dataset which has 5 points hence 10 output units.I used grayscale and normalized image as input. Here is my ...
0
votes
0answers
28 views

Correct way of computing dice score for image segmentation?

In binary image segmentation, for given a set of images, it's true mask and predicted mask. How do you compute dice score? Should I compute the dice score for each image separately and then find mean ...
0
votes
1answer
31 views

best similarity measure for images with different angles

I want to compare different images (where the images are of the same setup but the angles with which the images are taken are different). I want to obtain some sort of similarity score. I tried using ...
0
votes
0answers
18 views

How metric learning works for content based item retrieval

I was doing some computer vision experiments and recently I have started learning about metric learning and the image retrieval problem. I was experimenting with the inshop image retrieval dataset to ...
0
votes
0answers
11 views

What is the difference between a bounding box and ROI (Region of Interest)

I was reading about the Fast RCNN for object detection. From what I understand, it uses pre-computed ROI's (using selective search) and uses these to predict the bounding box offsets and uses smooth ...
0
votes
1answer
29 views

How to model the probability of detecting an image, given it is seen multiple times

Are there any existing methods/models describing the probability of an object being detected by a computer vision algorithm given it is seen $n$ times at similar angles and orientations? I know that ...
0
votes
0answers
19 views

How to evaluate pix2pix?

As far as I know, to evaluate synthesized images it is proposed to use: human scoring, "Inception score", where in the second case the quality is rated based on a pre-trained Inception ...
0
votes
1answer
31 views

How to interpret fast-rcnn metrics?

I'm following this tutorial to fine tune Faster RCNN model, during training process a lot of statistics are produced however I don't know how to interpret them. what are major characteristics to look ...
0
votes
0answers
12 views

Faster RCNN is not able to recognize composed elements

I tried to fine tune Faster RCNN to realize object detection task with bounding boxes where I shall recognize face, eyes and mouth on image. However Faster RCNN seems to fail to recognize objects when ...
1
vote
1answer
22 views

What is the upscaling factor in super resolution with deep learning? [closed]

I have been reading papers on single image super resolution (SISR) and I frequently encounter X3 upscaling factor, X4 upscaling factors. Example: SRGAN mentioning x4 upscaling factor It would be ...
0
votes
0answers
43 views

How to convert RGB to One Hot encoding of Pixel in Pascal VOC Dataset?

I am trying to implement Semantic Segmentation on PASCAL VOC 2007 Dataset using Fully Convolutional Network. My Network outputs images of (Height, Width, Classes); but the training label masks are of ...
0
votes
1answer
284 views

Class token in ViT and BERT

I'm trying to understand the architecture of the ViT Paper, and noticed they use a CLASS token like in BERT. To the best of my understanding this token is used to gather knowledge of the entire class, ...
1
vote
1answer
47 views

How to estimate real distance between two detected objects in an image?

You may think this is a duplicate, but my situation is different than previously asked questions. The only information I have is the width and height of the bounding boxes of detected people. The ...
1
vote
0answers
18 views

How Does EAST detector implementation with VGG16 look? How many outputs does it have?

I was reading the Efficient and Accurate Scene Text Detector paper and saw the author reference VGG-16 as a possible stem "feature extractor" network. In the paper they say: In our ...
0
votes
1answer
157 views

Can somebody explain me the meaning of this sentence? (Color Similarity - Selective Search Algorithm)

This is a sentence from this article : Color similarity: Computing a 25-bin histogram for each channel of an image, concatenating them together, and obtaining a final descriptor that is 25×3=75-d. ...
0
votes
0answers
19 views

Predictions receiving in Unknown (8)[e,e,e,e,e,e,e,e] format from TensorflowJS Mobilenet

I have trained a mobilenet on diamond images to count diamonds in broswer. I then converted the SavedModel format to TFJS format. I have the following code in my JS file. ...
0
votes
0answers
8 views

Retrieving image masks from PASCAL and COCO segmentation annotations

How do I retrieve image masks for instance segmentation data which has annotations saved in the COCO and PASCAL annotation formats with python? The format is very confusing for instance segmentation ...
0
votes
0answers
57 views

Unable to Predict the custom trained mobilenet in browser using TensorflowJS

I am new to TensorflowJS and Javascript. I trained a mobilenet with images of diamonds using Tensorflow Object Detection so that it can detect Diamonds. The model is saved in SavedModel format. I ...
1
vote
1answer
27 views

Feature extraction from sequence of images with Siamese Neural Network

I am trying to train a neural network to recognize certain actions in short movies. Each such movie consists of a fixed number of frames, each frame - the image is of course the same size, after ...
0
votes
1answer
23 views

Training CNN: Understanding number data generated while training the model

I am training CNN on kaggle and my training and test datasets shapes are as follows: ...
2
votes
0answers
25 views

What is the difference in computational cost at inference time between object detection and semantic segmentation?

I am aware that YOLO (v1-5) is a real-time object detection model with moderately good overall prediction performance. I know that UNet and variants are efficient semantic segmentation models that are ...
1
vote
1answer
41 views

Transfer Learning on Resnets/VGGs — Validation accuracy can never be over 75%

I am trying to classify skin cancer images into two categories -- malignant and benign. Literatures suggest that using pre-trained resnet/vgg network achieves more than 90% accuracy. However, with my ...
1
vote
0answers
16 views

Machine learning model (neural network or SVM) for unequal feature matrices size

I have feature matrices obtained from visual bags of words model for various dictionary sizes. Example, Nx5, Nx10, …., Nx15000. Where N is the number of samples and 5, 10, …15000 are the visual ...
1
vote
0answers
20 views

Why are axes-aligned bounding boxes used in object detection

I understand (I think) why in object detection, the result is a rectangle: it is a simple shape that can be defined by 4 variables (2 pairs coords of opposite corners or 1 pair of coords + width and ...
0
votes
1answer
18 views

Which F1-score is used for the semantic segmentation tasks?

I read some papers about state-of-the-art semantic segmentation models and in all of them, authors use for comparison F1-score metric, but they did not write whether they use the "micro" or &...
1
vote
0answers
87 views

Yolo issue with detecting positives

I've recently tried to implement a Yolo detector for traffic light detection based on yolo v1 implementation in Tensorflow/Keras. My model really struggles with detecting small objects. Loss function ...
0
votes
0answers
6 views

How do you compare outputs from different augmentations for consistency training?

I'm trying to figure out how papers using consistency train on unsupervised data, and I'm stuck on how outputs from different augmentations are compared. Since the augmentations are transformations, I ...
0
votes
0answers
139 views

How to convert VGG VIA Instance segmentation annotation to COCO/PASCAL for Tensorflow Object Detection?

I have VGG VIA JSON annotations for instance segmentation for counting diamonds from a given image. The annotations are mixture of Circle, Polygons, Polylines. However, I am planning to use the ...
-1
votes
1answer
40 views

LED status detector [closed]

I have a piece of custom device that displays its working status by blinking LEDs fitted on the device. for example, the led is ON means the board is on led is OFF means the board is off. red led ...

1
2 3 4 5
10