Blog about Machine Learning and Deep Learning

I was very impressed after analysing how Convolutional Neural Network see the image. More about it in the link. As any model in Machine Learning makes some mistake, I would like to try to analyze why my model made that mistakes. In order to solve this question we can use... [Read More]

Analysis of results and playing with featuremaps In order to train model for activity classification I used pre-trained model, trained on ImageNet. The next step was to apply transfer learning. What is it? The main idea is to use weights that were trained on ImageNet and change them a bit... [Read More]

We have learned and read a lot about how does CNN work. However, it was interesting for me to understand why did the ConvNet decide to classify the image to that particular class. What elements did it relly on to chose the class. And actually how does the ConvNet see... [Read More]

Hello all who are interested in Deep Learning! I am happy to see you here! Today I will write about my experience with Convolutional Neural Networks! I will describe here how to get image description. The image description will be consist of two parts. First one is summary about objects... [Read More]

Class distribution by probabilities in all test dataset This file is written using R. Just because I love R and it is very easy to visualize any plots with ggplot library. The csv file is a dataframe, that we got in previous post in Part2 data_to_plot = read.csv("/home/veronika/materials/cv/detection/total_classes.csv") head(data_to_plot) ##... [Read More]

In this small post I decided to play more with SSD. My goal was to count all objects per all classes that Network was able to detect. Also, it was interesting how class distribution changes per different Neural Network’s certainty. import numpy as np import matplotlib.pyplot as plt import pandas... [Read More]