Image-Classification
Let's Classify the Pics
The problem Image-Classification solves
We Know as a human, we can recognise the objects easily with our Eyes and recognise with our brain. Then, How can Machines do Identify the Objects, Images, How can they Identify and classify the objects or Images. The Solutions is "Neural-Networks". Yes, It was Possible to classify the Images with the help of Neural Networks. In this project I had developed a model that can classify the images with the help of some python Libraries like torchvision, seaborn, PyTorch... classification process is to categorize all pixels in a digital image into one of several land cover classes, or "themes". This categorized data may then be used to produce thematic maps of the land cover present in an image.
Challenges we ran into
For Classification, we need a large dataset. I was bit harder to load a larger dataset as we arer having low amount of GPU. It was necessary to classify every image in the dataset. We have taken an image data set containing nearly 0.4M Images. to perform certain Operation, we have to classify every image. Sometimes Kernal was becoming "DEAD-STATE" due to shortage of GPU. this will create some illingius and might be risky one too. to overccome this, we are working on some advanced python libraries like tensorflow & keras.
Technologies used