Questions tagged [image-classification]
For questions about image classification: a decision problem where an algorithm must decide to which class ('cat', 'chair', 'tree') an input image belongs.
732 questions
8
votes
1
answer
179
views
Problem of freeze metrics after first epoch
I encountered a problem with metrics fading after the first training epoch. During the first epoch, the model training proceeds normally. The loss metrics decrease, and the accuracy increases. The ...
- 81
4
votes
1
answer
71
views
Using a CNN classification model on images that contain a scale bar
I'm using a CNN classification model that I trained to identify phytoplankton classes from png images. The images in the training set do not contain a scale bar.
However, some of the datasets I want ...
- 115
7
votes
1
answer
179
views
Effects of resizing training images during preprocessing CNN classification model
I'm trying to train a CNN model to identify phytoplankton species from a training set. During preprocessing, the images are resized to 224x224, which seems to be stretching or compressing the object ...
- 115
0
votes
1
answer
70
views
when testing with shuffled data, accuracy is high, but when testing with unshuffled data, accuracy is low
To be clear, I shuffled my data when I trained it. It is only the testing data that I modified to be unshuffled, and found that accuracy tanks. (i also used the same data for training and for testing)
- 71
2
votes
1
answer
85
views
Recognize data point patterns from scatter plots
I'm going to perform data point pattern recognition. The ideal shape of the scatter plot looks like this.
Now, suppose I have about 10,000 different samples, the following 3 images are the "...
- 21
7
votes
2
answers
277
views
Combining image and tabular data for a binary classification task
I'm working on a binary classification task where the goal is to determine whether a tissue contains malignant cells
Each instance in my dataset consists of
a microscope image of the cell
a small set ...
- 331
8
votes
3
answers
540
views
Best CNN architecture for multiple aligned grayscale images per instance
I’m working on a binary classification problem in a biomedical context, with ~15,000 instances.
Each instance corresponds to a single biological sample (a cell), and for each sample I have three co-...
- 331
2
votes
0
answers
68
views
DensNet169 model accuracy not increasing on medical classification dataset
I am training an DensNet model on medical dataset which has gold standards as per annotation. After training i noticed accuracy is just 60%. Later i performed following changes but still no luck. ...
- 123
1
vote
0
answers
56
views
Overfitting Problem - Medical Image Classification with Transfer Learning (Keras)
Short version: 100% training accuracy, 75-79% testing accuracy
Long version:
I'm a data science noob and my project is to create an ensemble model of 3 to classify retinal fundus (eye) images to 6 ...
- 11
3
votes
2
answers
80
views
Pre-specified 1x3x3 edge filters for CNN
I am reading a preprint by Linse et al. that provides an image of pre-specified edge filters that are 1x3x3 for which the various constraints are $\sum_i w_i=0$ and $\sum_i |w_i|=1$. (the authors ...
- 398
2
votes
1
answer
200
views
Confidence levels and error rates in binary classification models
I am a beginner self-learning machine learning and I'm currently dealing with a binary classification problem. I made a binary classifier with a basic neural network and I did some experiments with ...
- 23
0
votes
1
answer
72
views
Using ResNet50 with SE block on imbalanced data - pytorch
I worked with a breast cancer ultrasound image dataset containing 432 benign cases, 210 malignant cases, and 133 normal cases. Initially, I used a pretrained ResNet-50 model, which yielded the ...
0
votes
0
answers
37
views
Why is my trained model not recognizing new images that are different from the test dataset?
I had been training a model to classify a bunch of different images by different labels, namely 2 at first, and if it was working well, I would add more labels to train. I've gotten good results from ...
- 101
0
votes
1
answer
68
views
My keras CNN model gives good predictions in 10/11 classes but missleads to the 11th class. What can I do to improve?
My project involves classifying printed numerical characters from real-life essays. My dataset includes 11 classes ('0' - '10'), with the label '10' representing the '/' symbol.
The issue is that ...
2
votes
1
answer
345
views
Validation accuracy can't increase above 70%
I am building a classifying model to predict images over 3 classes. The data is balanced, with 10.5k images for train ( 3.5k for each ), 3k validation images ( 1k each ).
I increased my ...
- 31